Electrical assembly, control method, control unit, and electrical system

Abstract

The application relates to an electrical assembly (1100), a control method, a control unit (700) and an electrical system. The assembly (1100) comprises a stationary component (1200), comprising a plurality of at least seven coil units (1210), wherein each of the plurality of coil units (1210) can be controlled in order to generate a variable magnetic field, an actuatable component (1300) which is movable relative to the stationary component (1200) and comprises a magnet arrangement (1310), wherein each individual one of the plurality of coil units (1210) is designed to generate the respective variable magnetic field in such a way that, through interaction of the respective variable magnetic field with the magnet arrangement (1310), forces and / or moments act on the actuatable component (1300) in multiple degrees of freedom, so that by controlling any combination of a fixed number of the plurality of coil units (1210), a position and / or movement of the actuatable component (1300) relative to the stationary component (1200) can be controlled in six linearly independent degrees of freedom, wherein the fixed number is smaller than the number of coil units.

Description

[0001] Berlin Heart GmbH

[0002] 1.224 PCT / P150349PC00

[0003] Electrical assembly, control method, control unit and electrical system

[0004] The application relates to an electrical assembly, a control method for an electrical assembly, a control unit for an electrical assembly and an electrical system with an electrical assembly.

[0005] The subject matter of the application is particularly advantageous in the field of pumping systems, especially systems for cardiac support and / or for supporting fontanelle circulation. It can also be used for fuel pumps, vacuum pumps, circulation pumps, and other types of pumping systems.

[0006] Outside the field of pump systems, numerous applications for the subject matter of the application are possible. For example, it can be used in electric motors, fans, compressors, mixers, turbines, centrifuges, position control and stabilization systems (such as reaction wheels, stabilizers for optical imaging systems, etc.) and in other areas.

[0007] The aforementioned systems have in common that they comprise an electrical assembly with a stationary component and an actuable component that moves relative to the stationary component, whereby the actuable component is actuated by means of magnetic interaction between the stationary and actuable components. For example, the stationary component could be the stator of an electric motor, and the actuable component the rotor of the electric motor (with corresponding applications for pumps, fans, compressors, mixers, turbines, centrifuges, and other electrically driven systems).

[0008] In principle, there are six linearly independent degrees of freedom for a relative position and / or movement of the stationary component and the actuable component (in particular, three rotational and three translational degrees of freedom with respect to three orthogonal Cartesian axes can be represented, but a different choice or representation of linearly independent degrees of freedom is also possible).

[0009] In many cases, it is desirable or necessary for magnetic forces to act between the stationary component and the actuable component with respect to all six degrees of freedom. Various approaches are known from the prior art for this purpose; for example, each degree of freedom can simply be controlled by a separate actuator (e.g., the coil of a lifting magnet or the motor phase of an electric motor). Examples with exactly six independently controllable phases for controlling six degrees of freedom can be found in US 2018 / 238411 Al or JP 2017 107191 A. US 2018 / 238411 Al shows an approach with six coils, each of which simultaneously represents a phase and thus also an independent actuator.The six-degree-of-freedom control arrangement shown in JP 2017 107191 A uses several coils in a single phase to control one degree of freedom, and thus in some examples has more coils than controllable degrees of freedom. However, several coils are connected to form a single phase, meaning they can only be controlled together and thus constitute one controllable degree of freedom. The number of phases corresponds to the number of controllable degrees of freedom. A disadvantage of this approach is that if any single phase fails, six-dimensional controllability is no longer possible. To ensure full controllability even in the event of a single failure, the phases would have to be implemented in duplicate, since the control effect of each phase is orthogonal to the other phases.This is clearly evident when considering the Cartesian coordinate system, but it also applies to all other possible definitions of the coordinate system, even if it consists of unit vectors which are not orthogonally aligned in the Cartesian orthogonal coordinate system.

[0010] The arrangement described in US 2014 / 077627 Al, with six phases, each with one coil, is capable of controlling a carriage in six degrees of freedom. The direction of action of the individual phases is not aligned with the Cartesian orthogonal coordinate system and also changes with the position of the carriage relative to the coils of the stator phases. However, for each carriage position, a coordinate system consisting of six linearly independent unit vectors can be found, in which each phase acts on exactly one unit vector. Therefore, the same applies to controllability in the first-fault case as for US 2018 / 238411 Al or JP2017107191A. Furthermore, JP2017107191A, US 2018 / 238411 Al, and US 2014 / 077627 Al show a limited operating range in all degrees of freedom, especially in the rotational degrees of freedom, when viewed in the Cartesian orthogonal coordinate system. This makes these systems unsuitable for continuously rotating drive systems.

[0011] In all these approaches, the available space is not used efficiently in either the stator or the rotor. Furthermore, the efficiency is potentially low, since only a portion of the magnetic hardware is used to generate a large actuator force in a single degree of freedom, with the remaining components acting as ballast or restricting the coil cross-section of the active actuator. However, a compact design is just as desirable as efficient operation and a long service life.

[0012] In certain applications, such as cardiac support systems or safety-critical systems, redundancy of the actuators with respect to their ability to control all degrees of freedom may be desirable in order to ensure sufficient operational safety, in particular at least first-fault tolerance (e.g., regarding the failure of a first actuator and / or associated components). This, in turn, can compromise compactness and / or efficiency. In the example described above, for instance, twelve actuators would already be required to implement each individual actuator redundantly.

[0013] Against this background, the application is based on the task of providing solutions regarding the magnetic storage and / or control of an actuable component relative to a stationary component that at least partially meet the aforementioned requirements or provide a suitable balance of the desired properties and / or at least partially avoid or reduce the aforementioned disadvantages.

[0014] To solve the problem, the subject matter of the independent claims is proposed. Preferred embodiments and optional further developments are defined by the features of the dependent claims. A disclosed electrical assembly (hereinafter sometimes also referred to as a machine) comprises a stationary component, comprising a plurality of coil units, wherein each of the plurality of coil units is controllable to generate a variable magnetic field, and an actuable component, which is movable relative to the stationary component and comprises a magnet arrangement.

[0015] It may be provided that the majority of coil units comprise at least seven coil units. It may be provided that the majority of coil units comprise exactly seven or eight coil units. The effects and advantages of such a choice of the number of coil units will become clear below.

[0016] Preferably, each of the plurality of coil units is arranged to generate the respective variable magnetic field in such a way that, through the interaction of the respective variable magnetic field with the magnet arrangement, forces and / or moments act on the actuable component with respect to several degrees of freedom, in particular such that by controlling any combination of a fixed number, for example at most six, of the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in six linearly independent degrees of freedom.The latter property can be designed such that by controlling any combination of a fixed number of the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in six linearly independent degrees of freedom, where the fixed number is less than a total number of the plurality of coil units.

[0017] The exact number of coil units required, as well as their wiring, arrangement, and shape, must be determined for each individual application. The application provides appropriate tools to facilitate this determination, both in general principles and with concrete examples.

[0018] The property that by controlling any combination of a fixed number, for example at most six, of the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in six linearly independent degrees of freedom, and / or that by controlling any combination of a fixed number of the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in six linearly independent degrees of freedom, where the fixed number is less than a total number of the plurality of coil units, is also referred to here as a selection property.

[0019] This selected feature allows the actuable component to be efficiently and compactly mounted and moved (e.g., rotated, translated, and / or tilted) in all degrees of freedom relative to the stationary component. The efficiency advantage arises from the fact that multiple, and in particular all, selected coil units can be used for actuation in any direction. Distributing the load across multiple actuators reduces the stress on each individual actuator. Since actuator power loss increases quadratically with actuator stress, the total power loss decreases. This design also facilitates redundancy and thus operational safety (especially first-fault safety, and, if required, second-fault safety). This safety, combined with the comparatively simple design, also enables a long service life.

[0020] The property that each of the plurality of coil units is configured to generate the respective variable magnetic field in such a way that, through the interaction of the respective variable magnetic field with the magnet arrangement, forces and / or moments act on the actuable component with respect to multiple degrees of freedom—that is, that each (especially independent) coil unit is capable of actuating the actuable component in multiple degrees of freedom—is referred to here as the universal actuator property. This property makes the actuators interchangeable. This also particularly benefits the selection property.

[0021] A given coil unit can comprise one or more coils or turns. A coil unit specifically includes those coils or turns which, due to their interconnection, can only be energized together, but independently of other coil units.

[0022] In the case of an electric motor, a coil assembly can correspond in particular to a motor phase. A fault against which the assembly according to the application provides protection can, for example, be a broken coil or a short circuit (an unintended electrical connection) between two coils, or a combination of such faults, or comprise them.

[0023] The property that by controlling any combination of at most six of the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in six linearly independent degrees of freedom can also encompass the case where controllability in six linearly independent degrees of freedom is already given by any combination of a given number of coil units, where the given number is less than six, for example, five or four. However, for quasi-static controllability (see below), a selection property with respect to six arbitrary coils will generally be required. This is because the information content of a predefined vector, for example, a coil unit current vector, with fewer than six entries is insufficient to uniquely represent six linearly independent actuator forces or actuator moments.

[0024] The actuable component can be rotated arbitrarily around a rotational axis relative to the stationary component. The actuable component is then also referred to as a rotor. The assembly is thus suitable for rotating systems such as pumps (including those for cardiac support), electric motors, fans, compressors, mixers, turbines, centrifuges, etc. It can be configured that, in any rotational position of the actuable component with respect to the rotational axis, the position and / or movement of the actuable component relative to the stationary component can be controlled in six linearly independent degrees of freedom by actuating any combination of the fixed number of coil units.

[0025] It can be designed so that the rotor's moments of inertia about several axes, especially all principal axes of inertia, are equal or exhibit only small deviations from each other, for example, less than 20 percent or less than 10 percent. This prevents the rotor from experiencing precession or nutation when the electrical assembly is tilted at high speed. This reduces the coupling of the degrees of freedom and simplifies or makes the control system more robust.

[0026] It can be provided that for each given coil unit of the plurality of coil units, a first position of the actuable component and a second position of the actuable component exist, such that in the first position, a force and / or a moment can be generated by interaction of the magnet arrangement with the variable magnetic field that can be generated by means of the given coil unit, which acts on the actuable component with respect to a first degree of freedom of the six linearly independent degrees of freedom, and

[0027] so that in the second position, through interaction of the magnet arrangement with the variable magnetic field that can be generated by means of the specified coil unit, a force and / or a moment can be generated which acts on the actuable component with respect to a second degree of freedom of the six linearly independent degrees of freedom.

[0028] In this way, each coil unit contributes to the actuation of the actuable component in at least two, or even more, degrees of freedom (across all possible relative positions of the stationary component and the actuable component), which in turn can contribute to the compactness, efficiency, and / or safety of the arrangement. This property is suitable as a criterion for the isolated design, layout, and optimization of the individual actuators, consisting of coil units and magnet arrangements. The exact number of coil units required, as well as their interconnection, arrangement, and shape, must be determined for each individual application. The application provides corresponding tools that enable this determination, both in general principles and with concrete examples.

[0029] Similarly - and in a formulation independent of a choice of coordinate system - it can be provided that for each given coil unit of the plurality of coil units, a first position of the actuable component and a second position of the actuable component exist,

[0030] so that in the first position, a first force-torque vector (also force-moment vector, see below) can be generated by the interaction of the magnet arrangement with the variable magnetic field that can be generated by means of the given coil unit, which acts on the actuable component, and so that in the second position, a second force-torque vector can be generated by the interaction of the magnet arrangement with the variable magnetic field that can be generated by means of the given coil unit, which acts on the actuable component.

[0031] where the first and second force and moment vectors are linearly independent of each other. The force and moment vectors can be represented, for example, in Cartesian coordinates (x, y, z) and also as forces. x , y , z The force and moment, x, y, z vectors are denoted. It can be provided that for each of the six given degrees of freedom, or more generally for each given system of six linearly independent degrees of freedom, and for each of these six degrees of freedom, at least one first coil unit of the plurality of coil units and one second coil unit of the plurality of coil units exist, such that by interaction of the magnetic arrangement with the variable magnetic field that can be generated by each of the first coil units and the second coil unit, a force and / or a moment can be generated that acts on the actuable component with respect to the given degree of freedom.

[0032] In this way, multiple coil units contribute to the actuation of the actuated component in each degree of freedom, which in turn can contribute to the compactness, efficiency, and / or safety of the arrangement. In particular, redundancy is thus provided for each degree of freedom, and therefore for the entire system. While this design requirement can also be met by actuators with single redundancy, in conjunction with the aforementioned design requirement that each actuator is capable of handling multiple degrees of freedom, a comprehensive design requirement for individual actuators is provided, from which particularly compact and efficient bearing systems can be assembled. The exact number of coil units required, as well as their interconnection, arrangement, and shape, must be determined for each individual application.The application provides appropriate tools that enable this determination, both in general principles and by means of concrete examples.

[0033] To achieve full redundancy with just one additional actuator, such as an additional coil or phase, each actuator must be designed to act on multiple degrees of freedom. This must also hold true for all possible definitions of degrees of freedom, regardless of the specific definition used. This includes coordinate systems where the translational degrees of freedom are not perpendicular to each other, coordinate systems where the axes of the rotational degrees of freedom are not perpendicular to each other, and coordinate systems where the degrees of freedom consist of linear combinations of arbitrarily angled translational and rotational degrees of freedom.

[0034] Each of these coordinate systems with 6 degrees of freedom can be mapped in the conventional Cartesian coordinate system. Each degree of freedom can be assigned a 6-dimensional vector, which corresponds to the mapping in the Cartesian coordinate system. To ensure that the coordinate system is 6-dimensional, it is sufficient that the matrix consisting of the 6 6-dimensional vectors of the different degrees of freedom has a rank of 6.

[0035] To achieve 6-dimensional redundant controllability with only one additional actuator, it must therefore be the case that in every 6-dimensional coordinate system each degree of freedom can be controlled by more than two actuators.

[0036] Each of the multiple coil units can be interconnected in such a way that it can be controlled independently of each other of the multiple coil units to generate the respective variable magnetic field.

[0037] Such a circuit can, in particular, contribute to or ensure the aforementioned properties regarding the controllability of degrees of freedom by means of the coil units in a simple manner, possibly together with a suitable choice of the number, arrangement, and / or shape of the coil units as described in this application. Independently controllable in this context means that any combination of phase currents can be implemented. This may also require an adjustment of the control of other phases. However, if a phase current is predetermined by other phase currents, then it is not independent. In the usual three-phase star connection with active control of the three phases and without active control of the neutral point, only two independent phase currents exist, since the third phase current results directly from Kirchhoff's law.This constitutes a three-phase system, however, only two independent phases or coil units exist within the meaning of this application. In principle, the dimension of quasi-stationary controllability is limited upwards by the number of independent coil units or independent phases. The stationary component and the actuable component can form an electric motor, in particular an axial flux motor, or a component, in particular an axial flux component, of an electric motor. Each of the multiple coil units can correspond to a motor phase of the electric motor or the axial flux component.

[0038] An axial flux component, in this context, refers specifically to the case where an electric motor incorporates both radial and axial flux components (for example, referred to as a radial-axial motor, Raxial motor, or Radax motor); in this case, the axial flux component refers to the axial flux components. In another example, an axial flux component can also be formed by the preload in a combination of two preloaded radial flux motors.

[0039] The reason why the axial flux components in particular allow control in six degrees of freedom becomes clear when the controllability of a conventional radial flux motor is analyzed.

[0040] The forces transmitted to the centrally aligned rotor of a radial flux motor always act on the rotor in a radial (passing through the axis of rotation) or tangential (not intersecting the axis of rotation) direction for all coils of the radial flux motor, but never in an axial direction. Thus, all generable force vectors lie in a single plane. Even from any number of force vectors in a plane, only net forces in that plane (i.e., two-dimensional) and torque perpendicular to the plane (i.e., one-dimensional) can be generated. Therefore, the controllability of such a pure radial flux motor is limited to three dimensions.

[0041] The method and preferred embodiments will be explained below using axial flux motors as an example. However, these include all motors with axial flux components, such as axially preloaded radial flux motors.

[0042] The plurality of coil units can comprise a first group of coil units and a second group of coil units, wherein the coil units of the first group are arranged in a first plane to form a first axial flux motor stator, and the coil units of the second group are arranged in a second plane to form a second axial flux motor stator. The groups of coil units preferably act on the same actuating component. This component may include separate magnet arrangements for each of the groups of coil units. However, the groups of coil units can also act on both sides of the same magnet arrangement to keep the actuating component lightweight and compact and to utilize the magnetic material particularly efficiently.Such arrangements are known for three-phase axial flux motors, but in the known symmetrical configurations they can only generate torques and not axial forces, and are therefore only controllable in one dimension. The individual groups of coil units do not necessarily have to be controllable in six degrees of freedom on their own; instead, full controllability can also be achieved in combination.

[0043] In advantageous examples, the first axial flux motor stator and the second axial flux motor stator each have exactly three coil units, and the magnet arrangement has one, three, or five pole pairs.

[0044] The electrical assembly may be designed with a design-related offset between the respective directions of action and / or coil axes of the coil units of the first and second axial flux motor stators. During operation, this design-related offset can result in a phase shift in the control signal. As explained below, this can offer advantages for controllability. The direction of action is defined here as the direction in which the respective coil unit is oriented to generate a force and / or torque on the magnet assembly.

[0045] The direction of action can refer to the effect of a force acting on the actuable component at a defined point of application with a defined direction. Additionally or alternatively, the direction of action can refer to torques about defined axes of rotation. The offset can be defined with respect to the points of application, the directions of the forces, or the axes of the torques, but is not limited to these. In general, the offset is intended to prevent the contribution of the coil unit or phase to the actuation, represented as a force-torque vector, from being linearly dependent on the contributions of the other phases or coil units, and thus prevent a degradation of controllability.

[0046] The design-related offset can be achieved in particular by a rotational offset of an arrangement of the coil units of the first axial flux motor stator relative to an arrangement of the coil units of the second axial flux motor stator, and / or by a rotational offset of respective pole pairs of two magnet arrangements relative to each other, and / or by an oblique magnetization of a common magnet arrangement. It can be provided that by controlling any combination of at most five, in particular at most four or at most three, of the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in the six linearly independent degrees of freedom, utilizing an inertia of the actuable component in a moving, in particular rotating, state of the actuable component.

[0047] This type of controllability, which utilizes the inertia of the actuable component in its moving state to control all relevant degrees of freedom, is also referred to here as dynamic controllability. This is distinct from quasi-static controllability, where all relevant degrees of freedom can be controlled even when the actuable component is at rest relative to the stationary component.

[0048] A quasi-static controllability in all six linearly independent degrees of freedom can be given, in particular, on the basis of the above-mentioned selection property with respect to any six coil units.

[0049] In one example, quasi-static controllability based on the selection property with respect to six arbitrary coil units can be combined with dynamic controllability using a smaller number of coil units as mentioned above (i.e., in such a way that by controlling an arbitrary combination of at most five, in particular at most four or at most three, the plurality of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled by utilizing an inertia of the actuable component in a moving, in particular rotating, state of the actuable component in the six linearly independent degrees of freedom).This combination can be exploited in a control method that allows switching from quasi-static control during normal operation to dynamic control in a fault condition that restricts the functionality of at least one coil unit, as described below.

[0050] It may be provided that the majority of coil units comprises at most nine or at most eight, in particular exactly nine, exactly eight, exactly seven or exactly six coil units.

[0051] The selection property described above can be provided in a particularly compact, economical, and / or efficient manner by choosing the number of coil units accordingly (trivially in the case of six coils, since only a selection of six coil units is possible). A detailed analysis regarding the controllability in the case of different configurations with respect to the number and arrangement of multiple—especially independently connected—coil units is given below.

[0052] In an advantageous example, the plurality of coil units comprises exactly seven coil units, and the magnetic arrangement of the actuable component comprises exactly two, five, or nine pole pairs. In another advantageous example, the plurality of coil units comprises exactly eight coil units, and the magnetic arrangement of the actuable component comprises exactly two or six pole pairs, with an advantage being controllability in six degrees of freedom.

[0053] Regarding the magnetic arrangement of the actuable component, it may be provided that it comprises between one and nine magnetic pole pairs, in particular two, five, six or nine pole pairs. As shown in this application, for example, two, five or nine pole pairs are particularly suitable in combination with a stationary component with seven coil units, two or six pole pairs are particularly suitable in combination with a stationary component with eight coil units, etc., to achieve controllability in six degrees of freedom (see controllability analysis below).The fact that the majority of coil units comprises a predetermined maximum number or a predetermined exact number of coil units (here, at most eight, exactly seven, or exactly six coil units) means that, in addition to this maximum number or exact number of coil units that satisfy the selection property, the stationary component does not include any further coil units that also satisfy the selection property (together with the aforementioned coil units). This could mean, for example, that the stationary component does not comprise any further coils or coil units at all, but it could also include the case that the stationary component comprises additional coils with a different function (such as sensor coils, etc.) that do not satisfy the selection property together with the aforementioned coil units.

[0054] The electrical assembly can include a support device for supporting the actuable component in one degree of freedom such that the actuable component is passively and stably supported in this degree of freedom relative to the stationary component. The support device is preferably configured to assume a first and a second position, such that in the first position the support device engages with the actuable component for the aforementioned support, and in the second position it does not engage. Such a support device can, in particular, enable the determination of transverse dependencies or couplings between individual degrees of freedom, as described below. In the second position, the support device can no longer be part of the electrical assembly and can therefore be designed to be removable.

[0055] The stabilizing support provided by the support device may be sufficient to stabilize the actuable component to such an extent that it is passively supported in all degrees of freedom. In this configuration, the characteristics of the position sensors and / or the forces and moments generated by the coils can be measured. This measurement allows for the determination of control loop parameters, particularly the K inputs. m -matrix or the parameters of the degrees of freedom controllers, are determined.

[0056] The stationary component and / or the actuable component may include an acceleration sensor and / or a force sensor. This enables, for example, the detection and / or compensation of imbalances, as described below.

[0057] It may be provided that the magnetic field generated by a first of the plurality of coil units is asymmetrical, in particular with respect to rotation and / or reflection, relative to the magnetic field generated by a second of the plurality of coil units. It may be provided that the respective magnetic components, such as coils and / or stator teeth, of at least two coil units of the plurality of coil units are arranged asymmetrically to one another, in particular not equidistantly, and / or have different dimensions and / or shapes. Alternatively or additionally, magnetic components of the magnetic arrangement of the actuable component may be arranged asymmetrically to one another, in particular not equidistantly, and / or have different dimensions and / or shapes.

[0058] By generating fields asymmetrically in this sense and / or by using an asymmetrical / irregular arrangement and / or design of components, the controllability, in particular the reliability of the controllability, of the actuable component can be improved, as will be explained in more detail below.

[0059] It may be provided that each of the plurality of coil units is assigned to a respective sector of the stationary component, each of which comprises only a part, in particular an angular section, of the stationary component.

[0060] wherein one or more of the plurality of coil units comprise several coils which are arranged in and / or on the sector assigned to the respective coil unit and are interconnected in such a way that they can only be energized jointly, but independently of the other coil units of the plurality of coil units.

[0061] It can be provided that the sectors assigned to the coil units are evenly distributed across the angular circumference of the stationary component. The coils of the associated coil unit arranged in a respective sector can be evenly distributed across the sector's angular circumference. However, non-uniform arrangements are also possible in both cases. The sectors can be arranged in an overlapping manner. The coils of the associated coil unit arranged in a respective coil sector can then, for example, also be located between coils of other coil units.

[0062] A disclosed control method is suitable for controlling an electrical assembly comprising a stationary component with a plurality of coil units, each of which is controllable to generate a variable magnetic field, and an actuable component that is movable relative to the stationary component and comprises a magnet arrangement. In particular, the control method is suitable for controlling an electrical assembly of the type discussed above, which fulfills the selection property, in its various possible embodiments.

[0063] The tax procedure includes:

[0064] Determining a plurality of control parameters for controlling each of the plurality of coil units,

[0065] Controlling and / or regulating a position and / or movement of the actuable component relative to the stationary component by applying a signal to the plurality of coil units according to the specified plurality of control specifications.

[0066] It may be provided that the majority of control parameters are determined as a control parameter vector using matrix multiplication of a commutation matrix, in particular a commutation matrix dependent on the position of the actuable component, with a force-torque vector. The force-torque vector is a vector with force and / or torque entries that relate in particular to forces or torques that can be exerted on the actuable component by appropriately controlling the majority of coil units. The commutation matrix may be a predefined commutation matrix, for example, before normal operation. The aforementioned force-torque vector may be a target force-torque vector.

[0067] It may be provided that the commutation matrix is ​​determined based on a linkage matrix, taking into account at least one optimization criterion, for example, minimizing power loss and / or maximizing controllability, in particular by inversion or pseudo-inversion of the linkage matrix. The linkage matrix can be defined such that the force-torque vector (possibly up to computational or optimization errors) results from matrix multiplication of the linkage matrix and the control input vector. The linkage matrix can represent a mathematical model of at least a part of the electrical assembly, which determines an actual force-torque vector as a function of arbitrary control input vectors for different positions of the actuable component relative to the stationary component.

[0068] For illustrative purposes, it is assumed here that a closed-loop control system is used. In the actual controller, a (target) force-torque vector is determined for the forces to be generated, and then the control input is determined via the commutation matrix. The actual (actual) force-torque vector acting in the controlled system can be determined from the control input vector via the linking matrix. In the control loop, this corresponds to the controlled system; therefore, this calculation is not performed. A model of the controlled system is only used for the preliminary calculation of the commutation matrix. This model is characterized by the formula: Force-torque vector = (linking matrix) x (control input vector). In other words, the first part of this formula is calculated continuously during operation of the assembly. The second part is determined in advance, e.g., once based on simulations.

[0069] The commutation matrix can be determined using a simulation, for example, using the finite element method, and / or by measurement. Such a measurement can be performed, for example, using force and / or torque sensors on the actuable component. Alternatively, a measurement can be performed by moving the actuable component in different directions, particularly in its linearly independent degrees of freedom, and detecting the back EMF induced in the coil units by these movements.

[0070] It can be implemented that the commutation matrix is ​​determined as a lookup table for various positions of the actuable component, for example, rotational positions of the actuable component around one or more rotational axes, especially principal axes, and stored in memory. This ensures a high update rate of the control loop even if calculating the commutation matrix is ​​computationally intensive during runtime.

[0071] If movements in multiple or all degrees of freedom need to be considered, the lookup table can become very large, especially too large for typical microcontrollers. It can be implemented that a first lookup table, stored in memory, is used for the commutation matrix to control the position and / or movement of the actuable component in at least one degree of freedom, while simultaneously a second lookup table is being calculated, particularly for more precise control of the position and / or movement and / or control of the position and / or movement in one or more additional degrees of freedom. It can be implemented that, upon completion of the calculation of the second lookup table, the control system switches from the first to the second lookup table, for which the commutation matrix is ​​calculated.

[0072] It can be provided that the commutation matrix can be represented as at least a three-dimensional control formula for a motor with at least four phases. The implementation of a control formula as matrix multiplication is known for one- or two-dimensional controls within the framework of field-oriented control, where the two dimensions comprise torque and field weakening. A system can also have several of these two-dimensional field-oriented controls. The advantage of the three- to six-dimensional control formula over several two-dimensional control formulas is that, to correct couplings or to apply the selection criterion in the event of a phase failure, the electrical assembly can be easily corrected, since the system behavior is described by a matrix and can therefore be globally optimized with minimal effort.

[0073] The control procedure can include: determining a cross-dependency, i.e., a dependency of the control and / or regulation of the position and / or movement of the actuable component relative to the stationary component with respect to a first degree of freedom on the position and / or movement of the actuable component relative to the stationary component with respect to a second degree of freedom.

[0074] Determining the cross-dependency and / or determining components of a linkage matrix of the type defined above can include:

[0075] Applying a modulation signal to a control signal for the electrical assembly,

[0076] Capturing a measurement signal that includes a signal component corresponding to the response of the electrical assembly to the modulation signal, determining the cross-dependency and / or the components of the logic matrix based on this signal component.

[0077] The control signal to be applied can correspond to a target position of the actuable component in at least one degree of freedom and / or a force and / or torque specification (in particular controller output) for at least one degree of freedom of the actuable component and / or a phase voltage applied to at least one of the plurality of coil units and / or a phase current flowing through at least one of the plurality of coil units.

[0078] The measurement signal to be acquired can correspond to a position of the actuable component and / or a back-EMF (see above).

[0079] The control procedure can include: determining and / or adjusting the control parameters, taking into account the specific cross-dependency, in particular to minimize the cross-dependency. By determining the cross-dependency and / or accordingly determining and / or adjusting the control parameters, smooth operation of the assembly can be enabled, for example even in the case of manufacturing variations and / or component drift.

[0080] The control method may include: minimizing an imbalance of the actuable component with respect to a rotation relative to the stationary component about an axis of rotation by dynamically setting and / or adjusting a position and / or orientation of the axis of rotation with respect to the actuable component, in particular based on sensor data acquired by means of an accelerometer and / or force sensor arranged on the actuable component and / or stationary component and / or based on minimizing at least one of the entries of the force-moment vector.

[0081] Minimizing imbalance (also known as balancing), possibly including the corresponding acquisition of sensor data, can be performed over varying time periods, for example, over several or many rotations of the actuated component relative to the stationary component, but also over shorter periods, such as one rotation or less. The precise methods are known from five-dimensional magnetic bearings, which typically have individual magnetic bearings for each degree of freedom. Calculating the corrected axis of rotation or compensation forces based on actuator forces or position / acceleration sensor data is also applicable here. Methods for actively balancing multidimensional magnetic bearings are generally known from the prior art (e.g., US9479035B2). These methods are also applicable to the present electrical assembly.For this purpose, the sensor position as well as the actuator forces can be decomposed into individual degrees of freedom using the methods described below and then balanced as separate actuators. This allows all one- or two-dimensional balancing methods or disturbance force compensation methods (e.g., EP4249039A1) to be applied individually to each degree of freedom. In addition to imbalances, operation at high speeds due to rapidly changing control parameters and phase currents also presents a challenge. Therefore, a further development of the control method is described below, which enables position control even at high speeds.

[0082] Determining the plurality of control parameters can include: determining a respective phase current through each of the plurality of coil units. Determining the plurality of control parameters can then further include: transforming the determined phase currents into components of a degree-of-freedom current vector that correspond to a respective displacement of the actuable component in the direction of the axes of a coordinate system fixed with respect to the stationary component and / or a respective rotation of the actuable component about the axes of this fixed coordinate system.

[0083] Determining the plurality of control parameters can further include: determining a respective transformed control parameter, in particular a current parameter, for each of the coordinates of the fixed coordinate system using a respective controller.

[0084] The contributions of the degree-of-freedom current vector can then be compared with the corresponding contributions of the current specification, and thus the contributions of a degree-of-freedom current error vector can be formed.

[0085] Based on the individual current error components, the currents that lead to actuation in the selected degrees of freedom can be controlled individually and independently of the other degrees of freedom. The controlled currents do not typically exist one-to-one in the stator phases, but rather preferably represent meaningful combinations of phase currents that depend on the magnetic flux of the magnet arrangement contained in the actuable component relative to the stationary component. This orientation of the currents to be controlled towards the actual magnetic flux allows the individual current controllers to be decoupled from the rotor movements, especially high-speed rotation.

[0086] Determining the plurality of control specifications can involve: back-transforming the transformed control specifications into a stator-fixed (i.e., stationary with respect to the stationary component) reference system to determine the plurality of control specifications, in particular phase voltages or phase currents.

[0087] A control system with transformation and inverse transformation as described can be understood as a multidimensional variant of field-oriented control and is suitable, for example, for rapidly rotating motors where a phase shift between setpoint and actual currents can occur due to rapidly changing control parameters, such as in controller output stages.

[0088] The control method can include: determining the acceleration and / or rotational speed of the actuable component, and calculating the corrective forces and / or torques that must be applied to the actuable component to maintain it at a constant position relative to the stationary component. The control method can also include driving multiple coil units such that the actuable component is subjected to the calculated corrective forces and / or torques by the generated variable magnetic field. The corrective forces or torques can therefore be introduced into the control loop, particularly as feedforward control.

[0089] The control method may include: acquiring at least one measured variable that enables the detection of a fault condition of the electrical assembly, wherein the fault condition includes in particular a failure of at least one first coil unit of the plurality of coil units and / or an unintended electrical connection of at least two coil units of the plurality of coil units.

[0090] Detecting such a fault condition allows, for example, the issuance of a warning and / or the termination of operation of the electrical assembly, but also the adjustment of operation to maintain operation despite the fault condition. In this way, reliability and / or safety during operation can be improved.

[0091] When detecting the fault condition, it may therefore be possible in particular to: adapt a rule for determining the majority of control specifications in such a way that the control and / or regulation of the position and / or movement of the actuable component relative to the stationary component is carried out using the majority of coil units, taking into account the fault condition, in particular without the first coil unit.

[0092] For example, it may be provided that, after adapting the regulation by controlling any combination of at most five of the majority of coil units, a position and / or movement of the actuable component relative to the stationary component can be controlled in the six linearly independent degrees of freedom by utilizing an inertia of the actuable component in a moving, in particular rotating, state of the actuable component.

[0093] In particular, it is possible to switch from quasi-static control during normal operation to dynamic control when a fault condition is detected.

[0094] A disclosed control unit is configured to control an electrical assembly, in particular an assembly of the type described above in its various configurations. The control is preferably carried out according to one or more configurations of the control method described above.

[0095] A disclosed electrical system comprises an electrical assembly, in particular an assembly of the type described above in its various configurations, and a control interface by means of which the assembly, in particular the stationary component of the assembly, can be connected or is connected to a control unit, in particular the control unit mentioned above.

[0096] The electrical assembly can – as already mentioned in the examples above – form a rotary machine or part of a rotary machine, in particular an electric motor or part of an electric motor, wherein the stationary component is a stator of the electric motor and the moving component is a rotor of the electric motor. In this case, the electrical system is in particular a motor or rotary machine system. The rotor can be magnetically supported or mountable relative to the stator in at least one degree of freedom, in particular in all six linearly independent degrees of freedom, and in particular in an actively magnetically supported or mountable manner. The rotor can also be mechanically and / or hydrodynamically supported or mountable. A combination of several bearing types (mechanical, hydrodynamic, actively magnetic, passively magnetic, etc.) can be provided for different degrees of freedom (or even working together for one degree of freedom).

[0097] The electrical system can be configured as a pumping system for a gaseous and / or liquid medium. Besides pumps that transfer a medium from one location to another, such as fuel pumps or vacuum pumps, this also includes, for example, circulation pumps that mix the medium within a reservoir.

[0098] For example, magnetically levitated pumps are used where frictional seals, lubricants, or ball bearings are disadvantageous, such as due to requirements regarding contamination, friction performance, vibration, or chemical compatibility. However, magnetically levitated pumps are more sensitive to movement because large active forces are required to keep the rotor moving in sync with the stator. To mitigate this disadvantage, the rotor can be designed with buoyancy compensation. This means that the average density of the rotor is adjusted to the density of the pumped medium by means of cavities or additional mass.

[0099] It can therefore be provided that the actuable component is set up to convey a fluid, in particular by rotation relative to the stationary component, and that a mean density of the actuable component is adapted to a density of the conveyed medium, in particular by cavities and / or mass elements arranged in and / or on the actuable component.

[0100] If the rotor's center of gravity coincides with the center of lift of the displaced fluid, then accelerations and rotational movements are also balanced. However, unbalanced gyroscopic moments may still occur under certain circumstances and can be actively compensated for.

[0101] Buoyancy compensation of the actuated component can also be advantageous for particularly low-vibration drives. This drive balancing reduces the static forces necessary for levitation. If these forces are sufficiently small, passive magnetic components such as return irons or passive magnetic bearings can be omitted. In this case, the forces generated by the coil units are the only magnetic forces acting on the rotor. Since these forces are precisely known and can be precisely adjusted, very accurate active balancing is possible. For this purpose, buoyancy compensation can be used in combination with other measures that reduce rotor force, such as support and cover plates, double volutes, or a double-sided axial inlet for radial pumps.

[0102] The electrical system can be configured as a cardiac support system and / or fontanelle circulation support system, wherein the stationary component is a stator of a rotary fluid pump for pumping blood (rotary blood pump or simply blood pump), and the actuable component is a rotor of the rotary fluid pump. Reliability (especially first-fault tolerance), compactness, and efficiency are of particular importance for blood pumps. Furthermore, controllability (especially in the form of a control system) in all six degrees of freedom offers particular advantages in this context with regard to the smooth operation of a blood pump (both due to the good damping provided by active magnetic bearing support of the rotor in all degrees of freedom and due to the possibility of compensating for imbalances as described).

[0103] The electrical assembly can be configured as a reaction wheel, particularly for attitude control of a missile. The missile can be, for example, a satellite, an aircraft, or a spacecraft (including rockets, airplanes, and drones). A reaction wheel comprises a rotatable flywheel, which can be coupled to and / or formed by the actuating component (designed as a rotor). During operation, the rotor with its flywheel is accelerated or decelerated to accelerate or decelerate the satellite's rotation by conserving torque. The reaction wheel's rotor is typically supported by ball bearings. However, these have a limited rotational speed and / or lifespan and can fail prematurely due to electrostatic discharges or aging lubricant. Ceramic-based ball bearings have addressed these problems to date.However, magnetic bearing-based reaction wheels have also been developed, albeit with a separate motor and magnetic bearings for different degrees of freedom. A particular application of magnetically levitated reaction wheels is the compensation of micro-vibrations, which can be caused by the use of ball bearings or originate from other components in the missile. In particular, unlike ball-bearing reaction wheels, they can reduce or even actively compensate not only for rotational vibrations but also for translational vibrations.

[0104] Another advantage of actively magnetically levitated reaction wheels is that the axis of rotation can be shifted relative to the stationary component, in this case fixed to the missile. With ball bearings, the axis of rotation is fixed. Therefore, only the single, fixed component of the angular momentum can be influenced by the ball-bearing reaction wheel. To align the missile along multiple axes, several reaction wheels or a suspension of the reaction wheel in an active gimbal—i.e., multiple actuator elements—are required. To maintain a specific orientation or to perform minor alignment corrections, it is sufficient to tilt the axis of rotation of the rotating reaction wheel (if possible) or to change its rotational speed. For larger course corrections, the reaction wheel can first be stopped, then accelerated around a first axis, and subsequently decelerated.The missile will then realign itself around a first missile axis. Subsequently, the reaction wheel's axis of rotation can be aligned with a second axis of rotation, and the reaction wheel can be accelerated and decelerated in this configuration. The missile will then realign itself with a second axis of rotation. However, the position of the reaction wheel relative to the stationary component and the missile's total angular momentum remain unchanged. By repeatedly shifting the axis of rotation and accelerating and decelerating the reaction wheel in the direction of rotation, the missile can be oriented incrementally at will, even though the reaction wheel's axis of rotation has only been adjusted within a limited range. This allows for the elimination of multiple reaction wheels or ensures emergency operation even with failed reaction wheels.

[0105] The electrical assembly, as registered, can replace both the motor and all active magnetic bearings and also offers first-fault safety with low complexity or weight costs.

[0106] Other applications include fans, compressors, mixers, turbines, centrifuges, and other rotating machines. The bearings used in most such applications (as well as for pumps, see above), such as ball bearings, plain bearings, or hydrodynamic bearings, have unacceptable drawbacks for some specialized applications, such as imbalance, wear, contamination, or inability to run dry. The specialized applications that therefore rely on magnetic bearings can be divided into two different categories:

[0107] i. fully active magnetic suspension with one phase for each degree of freedom,

[0108] ii. Passive magnetic suspension with active magnetic suspension in a subset of degrees of freedom.

[0109] Key disadvantages of a conventional fully active magnetic bearing are complex redundancy concepts and inefficient use of installation space.

[0110] A disadvantage of partially passive magnetic bearings is that, with regard to the relationship between restoring force and design, a passive magnetic bearing behaves like a spring with a positive or negative spring constant. In conjunction with the rotor mass, this results in a spring-mass oscillator, which can oscillate, especially at the resonant frequency. This oscillator is damped by the fluid or additional damping elements such as eddy current dampers. Particularly with gas as the conveying medium, the damping is very small and the resonance amplification can become very large. Even if the machine itself is not operated near the resonant frequency, neighboring machines, or in mobile applications the movement of the machine itself, can still excite the resonances.Furthermore, damage caused by contact between the rotor and the housing often goes undetected, as the rotor position in this axis is frequently not measured, since this is not required for position control.

[0111] The disadvantages of both approaches can be addressed by using the patented electrical assembly. In this case, the bearing itself can be entirely integrated into the assembly, which can optionally be fail-safe. However, it can also be advantageous to support the motor with passive magnetic bearings. Here, individual degrees of freedom are magnetically supported either stably or unstably, with the unstable degrees of freedom either predominating in the sum of the spring stiffnesses according to Earnshaw's theorem, or the sum being balanced. Zero-force control can be implemented in all magnetically supported degrees of freedom to absorb the bearing forces with very low power loss. Likewise, all degrees of freedom can be actively damped, making the machine insensitive to movements or disturbances from adjacent machines.

[0112] For applications where low system complexity or low costs are paramount, a motor with 6 or even down to 3 independent phases can be used. In this case, control, and especially damping, is achieved using the underdetermined control method presented.

[0113] Systems with non-rotating electrical components also offer advantageous applications. The electrical system can be designed, for example, as a stabilization device, particularly for an optical imaging system. The optical imaging system can be, for instance, an imaging system of a camera, a telescope, an exposure system, etc. It can be designed, for example, to stabilize an image sensor, a mirror, an imaging lens or lens assembly, or another optical component in its position and / or orientation by appropriately controlling the actuable component.

[0114] The drawings described below illustrate the background, principles, and exemplary embodiments of the subject matter of the application. FIG. 1A shows, schematically and not to a predetermined scale, a perspective view of an electrical assembly, and FIG. 1B a top view of a part of the electrical assembly according to FIG. 1A.

[0115] FIG. 1C shows a longitudinal section through the section plane A--A of the electrical assembly according to FIG. 1A,

[0116] FIG. 2 is a diagram showing the number of controllable degrees of freedom for various electrical assemblies.

[0117] FIG. 3 shows the distribution of magnetic fields in relation to coil positions in an exemplary electrical assembly.

[0118] FIG. 4 shows a schematic diagram of a control scheme according to an example,

[0119] FIG. 5 visualizes the angular dependence of a commutation matrix.

[0120] FIG. 6 shows a schematic diagram of a control scheme according to another example,

[0121] FIG. 7A and FIG. 7B Top views of a part of the electrical assembly according to FIG. 1A in various fault states,

[0122] FIG. 8 a diagram showing the number of controllable degrees of freedom for various electrical assemblies in a fault condition, FIG. 9 a visualization of the effect of a fault condition on the angular dependence of a commutation matrix,

[0123] FIG. 10A to FIG. 10D Top views of respective parts of electrical assemblies according to further examples,

[0124] FIG. 11 shows an exemplary measurement result for the determination and reduction of cross-dependencies,

[0125] FIG. 12A shows a longitudinal section through an electrical assembly according to another example,

[0126] FIG. 12B shows a cross-section through one of the planes A--A, C--C in FIG. 12A,

[0127] FIG. 12C shows a cross-section through the plane B--B in FIG. 12A,

[0128] FIG. 13 is a diagram showing the number of controllable degrees of freedom for various other electrical assemblies.

[0129] FIG. 14 a cardiac support system with a blood pump (in longitudinal section) and a control unit,

[0130] FIG. 15A shows a longitudinal section through a blood pump of a cardiac support system according to another example.

[0131] FIG. 16 shows a longitudinal section of a fan,

[0132] FIG. 17 shows a longitudinal section of a stabilization device,

[0133] FIG. 18 shows a longitudinal section of an optical imaging system with a stabilization device according to another example.

[0134] Recurring and similar features in the drawings are identified by identical or similar reference numerals. These may be partially replaced by TI.

[0135] They will be omitted if the relevant features are already shown and described in another drawing, or if they are not mentioned with reference to a drawing.

[0136] The electrical assembly 1100 shown in FIG. 1A to FIG. 1C comprises a stationary component 1200 and an actuable component 1300 (the latter is shown transparent in FIG. 1A for better visibility of the underlying components).

[0137] The stationary component 1200 comprises a plurality of coil units 1210, each of which can be controlled to generate a variable magnetic field. The actuable component 1300 is movable relative to the stationary component 1200 and comprises a magnet arrangement 1310.

[0138] In the example shown, the actuable component 1300 is movably mounted relative to the stationary component 1200 in six linearly independent degrees of freedom and, in particular, is rotatable about a rotational axis 1400, i.e., actuable in the sense of an electric motor (for example, for a heart support system or another application disclosed in this application). The stationary component 1200 thus assumes the role of a stator, and the actuable component 1300 the role of a rotor. In the following, these components will often be referred to as stator and rotor, respectively, according to this example, although the discussion is applicable to many other examples as well.

[0139] The motor in the example shown is an axial flux motor (also abbreviated as AFM) or could, for example, form an axial flux component of an electric motor in the sense mentioned above.

[0140] Each of the plurality of coil units 1210 is configured to generate the respective variable magnetic field such that, through the interaction of the respective variable magnetic field with the magnet arrangement 1310, forces and / or moments act on the actuable component 1300 with respect to several degrees of freedom, such that by controlling any combination of at most six of the plurality of coil units 1210, a position and / or movement of the actuable component 1300 relative to the stationary component 1200 can be controlled in the six linearly independent degrees of freedom (i.e., the selection property defined above is satisfied).

[0141] In the example shown, the stationary component 1200 comprises seven coil units 1210 (numbered 1 to 7 in FIG. 1B), each of which includes a coil (in the general case, one or more coils) with winders 1211 (in the general case, with one or more turns each) and corresponds to seven motor phases of the electric motor. In this example, the coil units 1210 are arranged coplanarly and equidistantly around the axis of rotation 1400. The magnet assembly 1310 comprises five pole pairs 1311 in this example. In the following, coil and coil unit are sometimes used synonymously, as in this example, although the discussion is also applicable to other examples in many cases (e.g., with more than one coil per coil unit, etc.).

[0142] Each of the plurality of coil units 1210 is interconnected in such a way that it can be controlled independently of every other of the plurality of coil units 1210 to generate the respective variable magnetic field. In particular, the coil units 1210 are not interconnected via common, uncontrolled corner or star points, which would reduce the number of independently controllable (here also referred to as independent) coil units 1210.

[0143] The electrical assembly 1100 exemplifies the aforementioned selection criterion, which can, however, also be met by other arrangements. The following discussion will focus on the principles of a controllability analysis, based on which a suitably appropriate number, arrangement, shape, and / or interconnection of the coil units 1210 can be determined.

[0144] To find out which degrees of freedom in a machine can be controlled independently, the machine can first be described using a matrix equation as follows:

[0145] ~^Fxil ^Fxi2 C-Fxip ' Cpyil ^Fyi2 Cpyip 1 z Cpzil Cpzi2 C-Fzip = FM = TM x m • I p ^Mxil C-Mxi2 C-Mxip ^Myil ^Myi2 C-Myip

[0146]

[0147] -CMzil ^Mzi2 C-Mzip- This includes:

[0148] FM is a force-moment vector consisting of the forces F acting on the rotor with respect to a Cartesian coordinate system (with position coordinates x, y, z). x , F y , F z and the moments M x , M y , M z ;

[0149] l p a phase current vector with phase currents ii to i p ;

[0150] p is the number of phases;

[0151] T m(also written as Tm) a 6 x p matrix. Without loss of generality, the z-axis is chosen here as the main drive axis, M z This is therefore the main drive torque. In most motors, the rotation around the main drive axis is not limited and therefore does not need to be controlled for position control.

[0152] T m is generally dependent on the position and orientation of the rotor:

[0153] Tm = T m f(x,y,x, 0 X , Oy, 0z),

[0154] where T m _f is a generator function of the linking matrix with the parameters of the rotor position x, y, z, where x is the rotor position in the x-axis (y, z accordingly) and the rotor orientation 0. X , 0 y , 0 Z , with 0 X(Theta-x) represents the rotation about the x-axis (y, z accordingly). Here, "position" refers to the translational position and "orientation" to the rotational position; however, in general, the terms "position" or "location" can be used collectively for both translational and rotational positions in this application. For non-active magnetically levitated drive systems, the term "rotor position" is also often used. However, this generally refers to the rotational rotor position in the direction of rotation. For example, the rotor angle used for commutation is often described as the rotor position. Thus, the rotor position in many representations differs from the general six-dimensional rotor position presented here.

[0155] T m_f can be determined by simulation, for example using the finite element method, or by analytical calculation. For an analytical calculation of T m In some cases, it may be useful to make simplifications to the model. Possible simplifications include representing extended magnets as a single magnetic dipole or an aggregation of magnetic dipoles. Coils can also be approximated as electrically controllable magnetic dipoles or as a few conductors of zero thickness.

[0156] For the present analysis, it is assumed that the space of motion is in x, y, x, 0. X , 0 y is limited. Since the rotation around the z-axis, 0 ZSince the drive axis represents the entire angular range from 0° to 360°, it must be covered. For higher numbers of rotor pole pairs, the range can be limited to the period (0°-360°) / number of pole pairs, provided the pole pairs are rotationally symmetrical. For the limited degrees of freedom, it can initially be assumed that T m The position changes only slightly when shifted, as a preferred position also exists here. This simplifies the generator function to

[0157] Tm_f(x,y,z,0 x ,0y,0 z ) ~ T m _f(x=0, y=0, z=0, 0 X =O, 0 y =O, 0 Z ) = T m _f(0z). The controllability for each rotor position can be determined by calculating the matrix rank of the T m -matrix can be determined:

[0158] n D or = Minimum( n D oF(0z) ) = Minimum( rank( T m _f(0 z ) ) ), for all 0 Z, with nooF: number of independently controllable degrees of freedom (DOF stands for "degree of freedom").

[0159] For nooF = 6, the rotor is fully controllable. This means the position in x,y,z and the rotation are 0. X , 0 y , 0 Z They can be controlled individually or in any combination.

[0160] For nooF < 6, only nooF are different linear combinations of F. x , F y , F z , M x , M y and M z controllable.

[0161] For an exemplary three-phase axial flux motor, as known from the prior art, nooF = 2 results, since only F z and M z can be controlled independently. Since only two currents can be independently applied to a three-phase motor with a star or delta connection, the information content in the current vector is also only two-dimensional. Quasistatically, nD The information content of the current vector should never be greater than the information content of the current vector.

[0162] To investigate further axial flux motor topologies, a T m An f-model can be created, which can represent any number of coils and pole pairs. Based on this, a controllability analysis can generate an overview of the controllability for different combinations of coils and pole pairs, as shown in FIG. 2.

[0163] With three, four, or five independently connected coils, as shown in FIG. 2, the appropriate number of pole pairs allows for the simultaneous (quasi-static) control of the same number of degrees of freedom, for example, with one pole pair. However, with some combinations, such as four coils and two pole pairs, the controllability is limited and the number of simultaneously controllable degrees of freedom is reduced. This can be attributed to symmetries between the coils, pole pairs, and the geometry. This symmetry between the magnetic field and the coil positions ensures that the force-torque vectors KM, which can be generated by the individual coils, are nThe currents point in a similar or even the same direction and are therefore linearly dependent. This linear dependency can lead to a situation where, despite having six or more controllable currents in combination, fewer degrees of freedom are controllable. Simple redundancy can also arise between two coils. Simple redundancy means that one coil can be replaced by another, but the two coils combined do not allow for more actuator directions than a single coil. For example, the number of simultaneously controllable degrees of freedom is never greater than three if the number of coils corresponds to the number of pole pairs.

[0164] It is particularly noteworthy that for six equidistantly arranged coils, there is no pole pair configuration with 6 degrees of freedom. Here, the symmetry of the stator limits the controllability.

[0165] Seven is the smallest number of equidistantly radially distributed coils with 6 controllable degrees of freedom (for 2, 5, or 9 pole pairs). The configuration with 7 coils and 5 pole pairs corresponds to the electrical assembly shown in Fig. 1A, 1B, or 1C.

[0166] All quasi-statically fully controllable configurations are circled in Fig. 2 (as well as in Fig. 8 and Fig. 13) and are to be understood as particularly advantageous configurations within the meaning of the application.

[0167] The special feature of a 6-phase AFM, namely that it has only 5 independently controllable degrees of freedom despite having 6 independently applicable currents, will be illustrated here with a numerical example. For the example, the T m -Matrix calculated for an AFM with a rotor with 5 pole pairs, since full controllability with 7 coils is to be demonstrated subsequently for the same rotor.

[0168] The rotor is assumed to be parallel to the stator. The rotor magnets of the rotor's magnet arrangement are selected, for example, using a Halbach arrangement, such that the magnetic field (B-field) generated by the rotor magnets at the level of the coils varies cosinally in the axial direction and sinusoidally in the tangential direction. Since the individual coils are significantly smaller than the rotor, they are approximated here as current-controlled dipoles. This results in the relationship between the dipoles at the coil positions and the axial and tangential magnetic fields shown in FIG. 3 before calculating the forces and moments. When the rotor rotates, the magnetic fields shift, while the coils remain stationary.

[0169] The local axial and tangential magnetic fields can now be determined at the locations of the coils. Since the coil is modeled as a current-dependent dipole, a coil current generates an axial force proportional to the axial rotor field and a tangential force also proportional to the tangential rotor field at the points on the rotor in the immediate vicinity of the dipoles. These forces act on the coils and, in the opposite direction, also on the rotor.

[0170] From the six axial and tangential forces, the forces and moments about the x, y, and z axes related to the rotor's center of gravity can then be calculated. Depending on the application, the center of gravity can be the center of mass, a magnetic bearing center of gravity (or a combination of both). The machine under consideration here has no passive magnetic bearings; therefore, T m relative to the center of mass and calculated as follows:

[0171] Btanl ' 5lTl( <p CO 11) Bfanö ' sin^CPcoilö) Bfan! ' COS(<p CO 11) Bfanö ' COS(SPcoilö) Baxl Baxö ' Baxl ' SWl((Pcoill) Fc ' Baxö ' sin^CPcoilö) Fc ' Baxl ' COS(<p CO 11) ' Baxö ' COS(SPcoilö)

[0172] ' Bfanö F xl..6 Fyi..6 F

[0173] (Matrix entries: . f 1 " 6 ).

[0174] M xl..6 Myl..f>

[0175]

[0176] M Z1 "6Those present are:

[0177] c N Coil force per magnetic flux density in Newtons per Tesla;

[0178] Btani ■ Magnetic flux density of the i-th coil in the tangential direction in Tesla;

[0179] B axi : Magnetic flux density of the i-th coil in axial direction in Tesla; tPcoiii ■ Angular position of the i-th coil in radians;

[0180] r c Radius of a circle on which the coils or dipoles lie, in meters.

[0181] <p co 2, as well as the T m -Matrix entries F X 2 and F y Two coils are shown in FIG. 3 for illustration. The forces of the other coils are determined equivalently, and the axial forces point from the coil positions into the figure.

[0182] The Tm matrix is ​​therefore valid for this specific case:

[0183] / F x \

[0184] — FM — T m6coils • I p —

[0185]

[0186] - 0 0.38 -0.38 0 -0.38 0.375 0 0.22 0.22 0 -0.22 -0.22 0.5 -0.25 -0.25 0.5 -0.25 -0.25 h 0 -0.01 -0.01 0 0.02 0.01 i40.03 -0.006 0.006 -0.025 0.006 -0.006 it

[0187]

[0188] - 0 0.02 -0.02 0 0.02 -0.02 \i6 / The number of independently controllable degrees of freedom corresponds to the rank of the Tm matrix. This can be calculated numerically and yields 5, which is less than the size of the square T matrix used here. m -Matrix.

[0189] The controllability can be illustrated particularly clearly if one considers T m represented by a QR decomposition through the product of a square matrix and a right triangular matrix:

[0190] T m 6coils öcoils -0.5 0.25 0.25 -0.5 0.25 0.25 - 0 -0.433 0.22 0.0007 0.43 -0.21 0 0 -0.37 0.003 -0.0002 0.38 Qöcoils 0 0 0 0.05 -0.025 -0.02

[0191] 0 0 0 0 0.04 -0.04

[0192]

[0193] - 0 0 0 0 0 0 - The linearly independent rows can be directly read from the R matrix. By entering the non-zero value on the diagonal of R, the rank and thus the controllability is reduced.

[0194] If one sets T analogously m For the 7-phase AFM with the same rotor with 5 pole pairs (i.e., approximately the example shown in FIG. 1A to FIG. 1C), the following T results. m - and R-matrix:

[0195] T^coüs = deoils ■ deoils -0.5 0.11 0.45 -0.32 -0.31 0.45 0.11 - 0 -0.49 0.14 -0.09 0.35 0.19 -0.11 0 0 -0.17 -0.37 0.17 -0.1 0.46 Qlcoils 0 0 0 -0.11 0.02 -0.004 0.09 0 0 0 0 0.04 -0.006 -0.04

[0196]

[0197] - 0 0 0 0 0 0.04 -0.04- Here, all diagonal entries have a value significantly different from zero, and full controllability is therefore given.

[0198] Instead of QR decomposition, a singular value decomposition of T can also be used. m This process can be carried out. In doing so, one not only obtains the number of degrees of freedom, but also, with the eigenvectors, a tool for selecting the axes of freedom to be measured or controlled, particularly in the case of an underdetermined control system. The axes to be measured or controlled, or the axes of the coordinate system, do not necessarily have to coincide and can be chosen arbitrarily and converted into one another.

[0199] Further characteristics of the exemplary electrical assembly 1100 shown in FIG. 1A to FIG. 1C (and further arrangements of corresponding controllability) are as follows.

[0200] Firstly, it follows that for each given coil unit 1210 of the majority of coil units 1210, there exists a first position of the actuable component 1300 and a second position of the actuable component 1300,

[0201] so that in the first position, a force and / or a moment can be generated by the interaction of the magnet arrangement 1310 with the variable magnetic field that can be generated by means of the specified coil unit 1210, which acts on the actuable component 1300 with respect to a first degree of freedom of the six linearly independent degrees of freedom and

[0202] so that in the second position, by interaction of the magnet arrangement 1310 with the variable magnetic field that can be generated by means of the specified coil unit 1210, a force and / or a moment can be generated which acts on the actuable component 1300 with respect to a second degree of freedom of the six linearly independent degrees of freedom.

[0203] Secondly, it follows that for each of the six given degrees of freedom, at least one first coil unit 1210 of the plurality of coil units 1210 and one second coil unit 1210 of the plurality of coil units 1210 exist,

[0204] so that by interaction of the magnet arrangement 30 with the variable magnetic field that can be generated by each of the first coil unit 1210 and the second coil unit 1210 a force and / or a moment can be generated which acts on the actuable component 1300 with respect to the given degree of freedom.

[0205] The following section will discuss control methods for controlling an electrical assembly 1100, including the example discussed above. Such control methods can, in particular, be implemented in a control unit for controlling an electrical assembly of the type described, wherein the control unit is configured to control an electrical assembly of the type discussed here.

[0206] In general, the control procedure includes: determining a plurality of control parameters to control each of the plurality of coil units,

[0207] Controlling and / or regulating a position and / or movement of the actuable component relative to the stationary component by applying a signal to the plurality of coil units according to the specified plurality of control specifications.

[0208] FIG. 4 shows a control scheme that implements an example of this control method. Dashed lines indicate which elements belong to the controlled system 424 and the controller 425, respectively. The majority of control inputs are represented as a control input vector (phase current target vector 306) using matrix multiplication of a commutation matrix (K). m ) 314, which depends in particular on the position of the actuable component 1300, is determined by a force-torque vector (actuator vector 304). The commutation matrix 314 will be discussed in more detail below.

[0209] To control or regulate a motor, the T discussed above is used. m not directly suitable, as it involves the T mThe rotor forces and torques are calculated from the phase currents using a matrix. However, for rotor position control, target forces and torques are determined using controllers. Subsequently, the necessary phase currents for these target forces and torques must be calculated. For symmetrical T m -matrices can be directly used for this purpose as the inverse T m Matrix can be used:

[0210] FM = T m • I p

[0211] => Ip = T m 1 • FM,

[0212] with T m -1 : Inverse of T M .

[0213] For this, T must m It must be invertible. This requires a square matrix with full rank.

[0214] For non-square matrices, the inverse is not defined. However, this is the case for all motors with 7 or more independent phases. Does T mIf a sufficient rank of 6 exists there, then there can be multiple solutions to the equation.

[0215] = K m = -^j =

[0216]

[0217] give. Since T m 1= K m K determines how the phases are to be energized. m (also written as Km) is called the commutation matrix (also commutation rule or commutation function). At runtime, only matrix multiplication is required to calculate the phase current target vector for any rotor forces or torques:

[0218] / p = K m • FM.

[0219] To choose from the many possible solutions for K m To determine the optimal solution, it may be possible to calculate the commutation matrix based on the linking matrix T. mis determined taking into account at least one optimization criterion, for example minimizing power loss and / or maximizing controllability.

[0220] For phases with identical coil resistance, the power loss is

[0221] pp

[0222] P = i n 2 ■ R = R i n 2

[0223]

[0224] n=ln=l

[0225] with

[0226] P: Total power loss in all coils,

[0227] i n : Phase current of the nth phase,

[0228] R: Phase resistance.

[0229] For equal phase resistances, minimizing power loss requires minimizing the sum of the squares of the phase currents. This also corresponds to minimizing the geometric sum of the phase currents.

[0230] I l geo = J lZyjp n =l T n2 >

[0231]

[0232] and thus also the matrix-2 norm. The minimization of the 2 norm upon inversion of a non-square matrix is ​​known in numerical analysis as the pseudoinverse. This is represented as a superscript "+". K m This allows it to be determined for each rotor position.

[0233] K m (x,y,z, 6 X , 6 y , 6 Z ) = Pseudoinverse(T m (x,y, z,d x ,d y ,d z = T m + , and for the specific case presented here

[0234]

[0235] = T m (0 z -) + .

[0236] If one places K m (0z) for the axial flux motor with seven coils and 5 pole pairs, then a K is obtained for the entire angular period of 0°-360° / pole pairs=72° m as shown in FIG. 5.

[0237] K mIt can be stored directly in the motor controller's memory, for example as a lookup table. When simplifying the magnets or coils to magnetic dipoles, the behavior of each component of K m sinusoidal. Therefore, every K m -Corponents for any number of rotor angles can be represented by a sine curve with amplitude and phase shift, and the course of the commutation matrix over the commutation angle can thus be stored in a memory-efficient manner.

[0238] For minimizations that differ from the 2-norm, the rows of K can be used. m(corresponding to the rows in FIG. 5) are minimized individually for each degree of freedom and each rotor position using iterative methods. The "conjugate gradient" minimization method generates results close to the pseudo-inverse. To improve the convergence of the minimization method, information about the shape of each matrix component can be used via the rotor angle. In this example, the curve is sinusoidal with a defined amplitude and phase shift and is also mean-free. Since fewer parameters now need to be optimized with the optimization method, it is generally more stable.

[0239] Now we refer again to FIG. 4. To control a rotor 309 in six degrees of freedom, the rotor position in these degrees of freedom is measured by a rotor position sensor 313. The sensor vector 301 with the translational positions x, y, z and the rotational positions 0 X , 0y , 0 Z The orientation can then be fed to a controller 302, which simultaneously receives a target vector. The target vector 303 can specify positions, and optionally also velocities and accelerations. The task of the controller 302 is to determine target forces and torques (actuator vector 304) that move the rotor towards the target vector.

[0240] The controller 302 can be implemented as a combination of individual controllers for each degree of freedom. All common controller structures are suitable for the individual controllers, for example, PID controllers, lead-lag controllers, fuzzy controllers, adaptive controllers, or extreme value controllers. The controller can also be implemented as a multidimensional controller, for example, a model-based controller. Observers such as Luenberger observers or Kalman filters with state feedback can also be used here. Based on the actuator vector 304, consisting of target forces and moments, a commutation 305 is then performed, i.e., the phase current target vector 306 is calculated, in particular—as already described—by matrix multiplication with the commutation matrix K. m 314. The Matrix K mis dependent on the rotor position. Depending on the signals (sensor vector) 301 of a rotor position sensor 313, which measures a rotor position 312, Km is selected from a lookup table or recalculated. Each column of the commutation matrix K m It contains the unit current vector and the phase current components necessary to generate one unit of force or torque of the respective degree of freedom. The matrix multiplication scales all unit current vectors by the force or torque specification (actuator vector 304) and sums all components of a phase to calculate the components of the phase current target vector 306.

[0241] For quasi-static control of the rotor position, the phase current target vector 306 can be directly scaled and output as a voltage to the motor phases by a power stage (here: power amplifier 307), for example, by PWM-controlled H-bridges or half-bridges. The resulting phase current 308 depends on the phase resistance, the phase inductance, and the induced back EMF. The phase resistance limits the phase current 308 for a given phase voltage. If the phase inductance and back EMF are low, then pure voltage control of the current may be sufficient. The back EMF is low during slow movements (including low speeds). The inductance is particularly low in ironless windings; therefore, these motors are especially suitable for matrix commutation.

[0242] At medium speeds, the influence of the inductance can be reduced by feedforward control, in particular by a phase offset d0. zThe commutation matrix is ​​determined, and these factors are compensated for. It should be noted that in multiphase motors, the mutual inductance can be significant, meaning that a current change in one phase can affect other phases. In three-phase motors, the coils can often be arranged so that the mutual inductances are almost completely canceled out by connecting the coils in a star or delta configuration. This advantage is only present in multiphase motors with certain current combinations, but not generally. The influence of back EMF can also be compensated for by feedforward control, in this case derived from the measured rotor position and its rate of change. The output stage can also include phase current measurement and current control for each phase.

[0243] The phase currents 308 impressed into the phases of the stator 309 in turn generate magnetic fields 310, which interact with the magnetic fields of the magnet arrangement of the rotor 311 and thus exert forces and moments on the rotor 311. For ironless windings, the complete force generation can be calculated using the Lorentz force.

[0244] The difference between the magnetic driving forces or torques and the output forces or torques, as well as the rotor inertia tensor, determine the rotor movement.

[0245] The rotor position 312 is measured in all degrees of freedom to be controlled via the rotor position sensor 313, thus closing the control loop. FIG. 6 shows a control scheme that implements another example of the described control method. The motivation for this, as already mentioned above, is that current controllers in the final stage of high-speed motors are subject to rapidly changing input conditions. This leads to a delay time and thus a phase shift between the actual and target current. This problem was solved in the prior art for synchronous and asynchronous machines using field-oriented control.

[0246] This concept can be extended to control six degrees of freedom in an n-phase motor. Determining the plurality of control parameters involves: determining a respective phase current through each of the plurality of coil units and transforming the determined phase currents 408 into components of a degree-of-freedom current vector 419, which correspond to a respective displacement of the actuable component (rotor) along the axes of a coordinate system fixed relative to the stationary component and / or a respective rotation of the actuable component about the axes of this coordinate system. This transformation defines a coordinate system, which is referred to here as the magnetic flux-oriented reference system 420. The previously considered (non-rotating) coordinate system is referred to here as the stator-fixed reference system 421.The phase currents 408 measured by current sensors 416 are thus converted by a rotor position-dependent matrix 417 into a virtual current vector (detected degree-of-freedom current vector 419), where each current corresponds to the force or torque generation of one degree of freedom. The rotor position-dependent matrix 417 corresponds to T. m , possibly scaled column-wise or row-wise to convert forces or moments into currents.

[0247] Each of the six current components can now be individually controlled in the rotating reference frame 420. In other words, a respective transformed control input is determined for each of the coordinates of the rotating coordinate frame 420 using a corresponding controller (individual current controllers 403). The individual current controllers 403 are decoupled from the rotor movement. The degree of decoupling depends on which components of the position detected by the rotor position sensor 413 are affected at T.m and K m The feedback can be provided. For conventional field-oriented control, only the rotation about the drive axis is fed back. This can also be sufficient for control in multiple degrees of freedom. However, more or even all degrees of freedom can be fed back. This allows the current controllers 403, for example, to be decoupled from vibrations caused by imbalance.

[0248] The individual current controllers 403 each output a voltage for one degree of freedom (summarized in the degree-of-freedom voltage vector 404). These up to six degree-of-freedom voltages are then multiplied, similar to matrix commutation, by the rotor-position-dependent K. m Matrix converted into phase voltage specifications (phase voltage vector 406 = control specification vector) (back-transforming the transformed control specifications to determine the plurality of control specifications). K mThis may require scaling row by row. These phase voltage specifications (phase voltage vector 406) can then be applied to the phases by the output stage (power amplifier 407) (phase voltages 415). Individual current controls or feedforward controls in the output stage are usually no longer necessary.

[0249] In the case that the rotor position sensor 413 only outputs the rotation about the drive axis and three motor phases with two current degrees of freedom exist, the following results for T m the well-known Clarke Park transformation and for K mThe inverse Park-Clarke transform enables the familiar one- or two-dimensional field-oriented control. For radial flux motors, only one-dimensional control of the drive torque and field weakening (which also only affects rotation) are possible. In single-sided axial flux motors, field weakening causes the stator to repel the rotor, thus generating axial forces. With negative field weakening, i.e., field strengthening, the rotor can be attracted. The standard field-oriented control is therefore capable of controlling a single-sided axial flux motor two-dimensionally, namely in the dimensions Mz and Fz, with rotational axis z.

[0250] Gyroscopic effects, such as precession or nutation, cause the control, particularly of the rotation axis orientation, to become unstable at higher rotational speeds using the controllers for each degree of freedom (which are stable at low speeds). Precession, for example, leads to a tilting of the rotor around the x-axis when torque is generated around the y-axis. This coupling of degrees of freedom can be taken into account when designing the controllers for the individual degrees of freedom, especially through coupling between the controllers. Suitable methods for this are known in the prior art (e.g., doi:10.1109 / 87.531916).

[0251] Building on the controllability analysis explained above, we will now discuss operation in the event of a fault. The effects of various fault scenarios, such as overheating, cable breakage, electrical breakdown of wire insulation or semiconductor components, mechanical damage, mechanical overload, or defective solder joints, can be categorized as follows: open phase, partially or completely short-circuited phase, a short circuit between adjacent phases, an open-type half-bridge switch, and a closed-type half-bridge switch.

[0252] FIG. 7A shows, by way of example, the stationary component 1200 of the electrical assembly 1100 according to FIG. 1A to FIG. 1C, where a break in a coil unit 1210 (i.e., open phase) has occurred as a first fault 1401. FIG. 7B shows an eight-phase stator in which a first fault 1401 as in FIG. 7A and, additionally, a break in another coil unit 1210 has occurred as a second fault 1402.

[0253] To react to a fault condition, it must first be detected. The control procedure can include: acquiring at least one measured variable that enables the detection of a fault condition in the electrical assembly, whereby the fault condition specifically includes the failure of at least one of the first coil units of a plurality of coil units and / or an unintended electrical connection between at least two coil units of a plurality of coil units. This can be done by continuous voltage or current measurement in the phases. If the phases are not measured individually, the total current can also be measured. The assignment to the individual phases can be done via an observer or a signal individually imprinted and identifiable in each phase.

[0254] One advantage of the seven- or multi-phase 1100 module is that, at least in the event of a first fault, all degrees of freedom can still be controlled if the operation is adjusted accordingly (in a situation where this is no longer possible, for example a warning can be issued and / or the operation can be terminated).

[0255] Controllability can be analyzed as described above. Controllability with a failed phase is shown in FIG. 8 for various numbers of phases and rotor pole pairs, as well as an initially symmetrical coil arrangement. In this case, the seven-phase motor with 2, 5, and 9 pole pairs remains fully controllable even in the fault condition. Controllability is only degraded by the fault condition for motors with 5 or fewer independent phases. In the fault cases mentioned above, with the exception of a closed-circuit half-bridge, it is possible to de-energize the affected phase. In the case of a short circuit between two phases, it is sufficient to deactivate one of the affected phases. After deactivating one phase, the seven-phase axial flux motor behaves like a six-phase axial flux motor with an asymmetrical or discontinuous coil distribution (FIG. 7A).

[0256] The asymmetry ensures that all degrees of freedom can still be controlled with the remaining six phase currents. This requires that all components of the K m -The matrix must be redefined. When detecting the fault condition, the following is specifically provided: Adapting a rule for determining the majority of control specifications in such a way that the control and / or regulation of the position and / or movement of the actuable component relative to the stationary component is carried out using the majority of coil units, taking the fault condition into account, in particular without the first coil unit.

[0257] In the present special case with 6 remaining coils and 6 degrees of freedom, T m square, so that K mIt can be calculated directly using the standard matrix inversion. Further optimization, for example of power loss, is no longer possible because no dimension is available and the matrix inversion is unique. This results in 7 different K values. m -Matrices, one each for deactivating a defined phase.

[0258] FIG. 9 illustrates the difference between K by way of example. m Normally (7 coils) and Km3 for a failure of phase 3. Emergency operation has a reduced efficiency but still full rotor controllability. The units correspond to the units in Fig. 5 with N / A for the forces and Nm / A for the torques. The abscissas are labeled here with the mechanical rotor rotation about the z-axis. The change in the entries of the commutation matrix K m The graph plotted against rotation about the z-axis is only an example. It may be advantageous to use K mlikewise, the position of the actuable component can be varied in further degrees of freedom, in particular all degrees of freedom.

[0259] With more than seven phases, two-phase faults (as shown in FIG. 7B) can theoretically be tolerated. However, in the specific case of eight phases, two-phase faults can result in a symmetrical six-phase motor, which may only be fully quasi-statically controllable in five degrees of freedom (FIG. 7B). Therefore, while the symmetrical eight-phase motor is statistically more fault-tolerant, it is not generally two-phase fault-safe. The nine-phase axial flux motor is thus the smallest symmetrical two-phase fault-safe axial flux motor with controllability in six degrees of freedom. A further advantage is that three-phase motors with three or six coils per phase are already being mass-produced. These can be operated in nine-phase mode without changing the winding, simply by rewiring the individual coils.

[0260] With more than nine phases, reliability can be further increased. However, the added value diminishes because individual components, such as the control unit, can still be susceptible to first-fault failures. For applications requiring even higher reliability, it may be advantageous to couple two separate first-fault-proof stators to the same rotor. In further configurations, the same magnet arrangement can be used for this purpose, so the rotor does not need to be larger or heavier.

[0261] Once the fault is identified, the corresponding output stages can be switched off and the commutation matrix K m , and possibly also T m, can be adjusted accordingly. In principle, several matrices can be stored for this purpose. If the matrix entries are calculated at runtime, the calculation rule can be changed. For the above example of switching from 7 to 6 phases, adjusting the amplitude and phase for the generator function of each entry of K is sufficient. m -Matrix. This allows for a seamless response to the fault. Furthermore, by controlling all degrees of freedom, comprehensive diagnostic capabilities are available to assess the damage and its impact. Based on this, repairs can be prepared and downtime minimized.

[0262] If the electrical assembly has fewer phases than degrees of freedom to be controlled (or degraded controllability), then quasi-static control is not possible. m It then has the dimension fxp (f: number of degrees of freedom, p: number of independent phases):

[0263] ~Cpxil Cpxi2 Cpxip ' 'y Cpyil Cpyi2 Cpyip Fz c Pzil c Pzi2 Cpzip = FM = T m • I p = M x CMXH ^Mxi2 C-Mxip

[0264] ^-Myil C-Myi2 C-Myip

[0265]

[0266] z - C Mzil C Mzi2 C-Mzip-

[0267] For p < 6, not all components of FM are independently controllable. This does not necessarily mean that some components of KM are not controllable, but could also mean that the degrees of freedom are coupled.

[0268] Therefore, quasi-static control in f degrees of freedom is not possible. However, dynamic control (as defined above) may be possible. In the aforementioned control method, for example, it may be provided that, after adjusting the rule by controlling any combination of at most five of the majority of coil units, the position and / or movement of the actuable component relative to the stationary component can be controlled in the six linearly independent degrees of freedom by utilizing the inertia of the actuable component in a moving, especially rotating, state of the actuable component. Thus, it is possible, in particular, to switch from quasi-static control during normal operation to dynamic control upon detection of a fault condition.

[0269] This will be explained in more detail. If the rotor moves, for example, rotates, the controllability changes with respect to the movement. To enable control of the desired f degrees of freedom, two conditions must be met:

[0270] 1. The couplings (columns of the different T m -matrices) at different rotor positions must span a vector space that is f-dimensional.

[0271] 2. The rotor's movement must be fast enough so that the rotor's inertia averages out the rotor forces and moments.

[0272] To determine the possible corrective forces, the following procedure can be used:

[0273] 1. Determine rotor position,

[0274] 2. Determine the desired correction force-moment vector (taking into account inertia, passive bearing forces, load forces or gyroscopic moments),

[0275] 3. Determination of T m for the current position,

[0276] 4. Determination of the possible vector space defined by the columns of T m is stretched

[0277] 5. Determination of a possible correction force-moment vector that is as close as possible to the desired correction force-moment vector.

[0278] 6. Imprinting the currents for this possible correction force-moment vector.

[0279] It is also conceivable to modify the methods explained in connection with FIG. 4 and FIG. 6 and to weaken or switch off individual degrees of freedom that are difficult to control at the current rotor position.

[0280] Alternatively, model-based methods such as Model Predictive Control and / or state feedback can be used.

[0281] The control system can be so restricted by the methods that position control in certain degrees of freedom or linear combinations of degrees of freedom is no longer possible. In these cases, however, the control system can often still be used for damping. Passive couplings between degrees of freedom, for example through gyroscopic effects or passive magnetic bearings, can be helpful here, since damping one degree of freedom also dampens all passively coupled degrees of freedom.

[0282] It was mentioned above that controllability can be degraded by stator symmetry, as is the case, for example, with a six-phase stator with coil units 1210 arranged equidistantly around the axis of rotation (FIG. 10A). In the case of fault operation, it was also discussed how this degradation can be counteracted by an asymmetrical arrangement. This effect can also be utilized in normal operation or taken into account in the design of the electrical assembly.

[0283] In the examples shown in FIG. 10B to FIG. 10D, it is provided that the respective magnetic components of at least two coil units 1210 of the plurality of coil units 1210 are arranged asymmetrically to each other, in particular not equidistantly, and / or are dimensioned differently and / or shaped differently.

[0284] Thus, the six-phase stator can achieve full controllability by means of a gap between two coils (FIG. 10B). This arrangement corresponds to a seven-phase stator with one phase dropped. An exemplary alternative arrangement of a six-phase stator includes a gap between groups of two coils (corresponding to a nine-phase stator with three phases dropped at equal intervals).

[0285] To make better use of the available space, the shape of the coils can differ from each other (alternatively or in addition to the asymmetries / gaps mentioned in the arrangement) (FIG. 10C, FIG. 10D).

[0286] The asymmetry can be generated additionally, alternatively, or in addition to the arrangement or properties of the coils, also with differently shaped or asymmetrically distributed stator teeth to achieve full controllability. Specific aspects and configurations of the control method will be discussed below.

[0287] If the aligned axis of rotation is not located on the rotor's center of mass, the drive rotation creates an imbalance that transmits forces to the rotor and counterforces to the stator. Not only the center of mass plays a role here, but also the bearings of the axis of rotation relative to the rotor's principal axes of inertia.

[0288] To counteract the negative effects of such imbalances (vibration, stress), the control method can include: minimizing the imbalance of the actuable component 1300 with respect to a rotation relative to the stationary component 1200 about a rotational axis 1400 by dynamically defining and / or adjusting the position and / or orientation of the rotational axis 1400 with respect to the actuable component 1300. Methods for such automatic minimization of imbalance by dynamically selecting the rotational axis relative to the rotor are generally known. This applies both to imbalance minimization in one plane and in up to six degrees of freedom.

[0289] If the goal is to minimize not the rotor's movement but the stator's vibration, this can be measured using at least one accelerometer and / or force sensor mounted on the stator, and the axis of rotation adjusted accordingly. For imbalances or higher-order vibrations, a purely static positioning of the axis of rotation is insufficient. In these cases, compensating forces can be continuously measured and applied.

[0290] Since all rotor magnets are available for generating drive torque and for bearing support, the rotor in the presented assemblies can be particularly lightweight. This results in low imbalance forces and, in combination with the imbalance compensation method, in exceptionally small imbalance forces.

[0291] The Km matrix was theoretically determined in the preceding examples. This involves making certain assumptions about, for example, the rotor's magnetic field, the coil geometry, the iron geometry, and the symmetry of all components. In reality, there will always be a difference between these assumptions and the actual setup. If these are systematic deviations, they can be measured on exemplary samples for the present machine series. However, this does not account for process variations or aging. By measuring and controlling the rotor position with ideally six degrees of freedom, the electrical assembly can measure itself. With fewer degrees of freedom, the information content decreases, but an equivalent measurement is still possible.

[0292] It is therefore possible to determine and / or correct the commutation matrix using a measurement. Such a measurement can be performed, for example, using force and / or torque sensors on the actuable component 1300. Alternatively, a measurement can be performed by moving the actuable component 1300 in different directions, particularly in its linearly independent degrees of freedom, and detecting the back EMF induced by these movements in the coil units 1210. These possibilities are explained in more detail below.

[0293] Another proposed method for self-measurement and correction of the Km matrix assumes that the initial K mThe matrix is ​​already capable of quasi-statically controlling the rotor in the degrees of freedom to be regulated. In this state, signals are then applied to individual force or torque components. These signals, and the resulting forces and torques, cause the rotor to move. This movement is determined using the degrees of freedom of the rotor position sensor or the responses of the remaining controllers. The measured sensor signals are then analyzed for signal components that correlate with the applied signals. The correlation should be greatest in the targeted degree of freedom, otherwise the control system would not be stable. Correlations in degrees of freedom that do not belong to the specific excitation signal indicate a transverse dependency. Such a transverse dependency can arise, for example, if the sensor orientation, and thus the sensor coordinate system, does not match the stator coordinate system.

[0294] If the sensor assembly is rotated by 10° around the z-axis, for example, then the x- and y-force generation appear to merge. If force is to be generated in the x-direction, then the actual x-force is reduced to cos(10°) = 98.5%. At the same time, however, an unintended y-force is generated with sin(10°) = 17.4% of the intended x-force. The correlation then determines precisely this 17.4% lateral influence. From the preliminary K m The matrix knows which currents should actually generate this 17.4% y-force. These current components are then determined by the K m Subtracted from entries for the x-force.

[0295] If this process is iterated for all degrees of freedom, multiple times if necessary, then K adapts accordingly. m The method adapts to the actual conditions of the assembly. It effectively measures the sensor characteristics in relation to the characteristics of the actuators.

[0296] The adjustment can be repeated for multiple electrical or mechanical rotor angles as well as displacements in the other degrees of freedom. This allows K m to be adapted to the individual field shape of the rotor as well as the shape of the coils and iron components.

[0297] The estimate for a sensor rotation of 10° illustrates that the system is sensitive to cross-dependencies. The coupling of the controllers of all degrees of freedom further increases the disturbance in each controller. Therefore, appropriate compensation methods should not be considered a further development, but rather a prerequisite for safe operation in setups with certain tolerances or design simplifications.

[0298] The applied signals simply need to be identifiable within the sensor signals. Examples include sinusoidal oscillations, pseudorandom signal signals (PRBS), or a simple offset. If the signals are also distinguishable from one another, then multiple degrees of freedom can be corrected simultaneously. Even during normal operation of the assembly, the signals can be applied at a sufficiently small amplitude to monitor or correct aging or drift.

[0299] The measurement can also be performed when the rotor is passively held. For this purpose, the electrical assembly can include a support device to support the actuable component in one degree of freedom such that the actuable component is passively and stably supported relative to the stationary component in this degree of freedom. In this setup, a small-signal analysis of the transverse dependencies of the commutation matrix is ​​possible.

[0300] FIG. 11 shows the result of an example measurement. The rotor is suspended in the z-direction by a thread. Therefore, only Fx, Fy, Mx, and My are measured. By optionally counteracting the movement around the z-axis and measuring the force on the thread, Fz and Mz can also be determined and corrected; this was not done in the example.

[0301] Each plot in FIG. 11 shows the rotor position in a specific degree of freedom during actuator control (actuated degrees of freedom) in another specific degree of freedom. Control was applied once in the positive direction and once in the negative direction to suppress offsets. Each plot is labeled {"Imposed force or torque" -> "Sensor channel"}. In the plots where the actuator degree of freedom and the sensor degree of freedom coincide, the effects are, as expected, large and relatively independent of the commutation angle.

[0302] The km measurement can be corrected as follows:

[0303] Determine or read off the coefficients of x_a, y_a, tx_a and ty_a from a plot (as in FIG. 11) such that:

[0304] x_a{x,y,tx,ty} = Sensor values ​​(x,y,tx,ty) / A_x with Ax being the amplitude of the excitation in the experiment.

[0305] Setting up the system of equations for correcting the actuator Fx:

[0306] ty / •$x=l %ax Vax iA / ax tyax a (a <? °y = 0 %-ay yay tx tyay b fy $tx=O %atx y atx LA 'atx ty at xc b c \yty=O Xaty y aty LA 'aty tyaty. dd

[0307] & CS- 1 ' $x0 —

[0308]

[0309] with CS - CrossSensitivity Matrix and matrix entries [actor-dof] a [affecteddof]- The corrected K m -Matrix entry K m2 for the X axis, then

[0310] ^

[0311]

[0312] m2 ^mx ' a "f” ^my ' "f” ^mtx ' "f” ^mty ' with K m{DOF}: Current vector for generating force in {DOF = degree of freedom}. The calculation can be repeated for the other actuator degrees of freedom with different a...d.

[0313] Alternatively, the commutation matrix with compensated cross-dependencies Km comp can be calculated directly:

[0314] K m I comp = CS -1 • K m ltc .

[0315] As described above, the tax procedure therefore includes: determining a cross-dependency and determining and / or adjusting the tax specifications taking into account the determined cross-dependency in order to minimize the cross-dependency.

[0316] In the previous examples, the actuable component / rotor was located on one side of the stationary component / stator. However, the electrical assembly can include a second actuable component / rotor. This second actuable component can be controlled by the multiple coil units of the stationary component in the same way as the aforementioned actuated component. Axial flux motors with rotors on both sides are known from the prior art to increase efficiency.

[0317] For example, if one assumes T for a double-sided three-phase axial flux motor m If this is the case, then it can be seen that the drive torque Mz can be generated more efficiently. The generation of force F z However, the direction of change is opposite for the two rotor halves and can even cancel each other out completely. When the generation of F zIf high efficiency is still desired, but less efficiency is required, then one rotor half can be equipped with smaller or weaker magnets to reduce the coupling between the coil and the magnet. The drive torque generation then benefits from both rotor halves, while the power generation behaves as if there were only one rotor with the difference in the number of magnets.

[0318] As illustrated, the second rotor half increases the axial magnetic flux but decreases the tangential magnetic flux. The radial winding strands generate Lorentz forces perpendicular to the field. Without tangential magnetic flux components, no force can be generated in the axial direction.

[0319] Extending this to a six-phase or more motor, it becomes apparent that each individual coil can generate more tangential forces but fewer axial forces. Due to the arrangement of the coils and the combination of forces on the rotor, Mz, Fx, and Fy are increased in the asymmetrical rotor, while Fz, Mx, and My are weakened.

[0320] As shown in FIG. 12A to FIG. 12C, it can also be provided that the plurality of coil units 1210 comprises a first group of coil units 1210 and a second group of coil units 1210, wherein the coil units 1210 of the first group are arranged in a first plane A--A to form a first axial flux motor stator and the coil units 1210 of the second group are arranged in a second plane C--C to form a second axial flux motor stator (parallel to the first plane A--A).

[0321] Fig. 12B shows section AA of the first group of coil units. The second group of coil units can be constructed similarly to the first group and, in addition, be positioned rotated relative to it.

[0322] Fig. 12C shows a section in the plane B--B through the actuable component 1300 with a magnet arrangement with one pole pair.

[0323] The discussed analysis of the controllability of axial flux motors can also be used to evaluate such combinations of axial flux motors. Even the combination of two axial flux motors with three individual phases can result in six degrees of freedom of control with only six coils. One advantage of using two opposing axial flux motors is that the static attraction between the rotor and stator irons can be balanced, since the two stator irons pull in opposite directions. In turn, the stator irons, unlike air-core coils, increase efficiency.

[0324] To achieve redundancy here as well, one or both stators can be designed with four or more independent phases. The rotor can contain magnets for each stator or use the same magnets for both stators. The latter case is illustrated in FIGS. 12A to 12C.

[0325] In addition to the number of coils in the upper and lower stators and the number of pole pairs, further configuration parameters emerge during the controllability analysis. For example, the stators can be additionally rotated relative to each other, which, as shown in the right half of Fig. 13, can positively influence controllability. The rotor can be designed as a disk with magnets, with each side facing a stator. Alternatively, each stator can actuate its own rotor magnet disk. In this case, the number of pole pairs of the two rotor magnet disks can also differ, or the pole positions can be rotated relative to each other. The positive effect on controllability of rotating the pole positions in a two-rotor magnet disk configuration can also be achieved in a single rotor magnet disk configuration through skewed magnetization or a stepped magnet arrangement.

[0326] Crucially, the direction of action of the individual actuator forces of the first and second stators is different, particularly shifted across the drive rotation direction (corresponding to a phase shift in the control signal). This different direction of action can be achieved through measures such as stator rotation or skew magnetization, but is not limited to these.

[0327] For the double axial flux motor with one disk and optional stator rotation relative to each other, the number of independently controllable degrees of freedom is shown in FIG. 13. To the left of the ordinate, the configuration of a first stator (2001) and a second stator (2002) is shown. The abscissa primarily shows the number of rotor pole pairs, but is additionally shown in two groups with (2004) and without stator rotation (2003).

[0328] Twisting the stators is necessary to achieve full controllability with the 3+3 coil configuration. Different twist angles result in varying degrees of controllability. For a twist angle of 1 / 3 of the coil spacing, some of the 3+4 configurations are also fully controllable.

[0329] The advantageous configuration with 3+3 coils and one pole pair corresponds to the embodiment shown in Fig. 12A, 12B and 12C.

[0330] For some configurations with a total of 7 or more independent phases, full controllability can be maintained even in the event of a fault. The commutation rule for the various fault conditions can be determined analogously using the same methods as those used to derive the commutation rule for the single-sided axial flux stator.

[0331] The configuration with 6+6=12 independent stator phases is known from the prior art (JP2010279230A) for controllability in six degrees of freedom; therefore, Fig. 13 only shows new configurations with a smaller number of independent stator phases in comparison.

[0332] The static twisting also allows quasi-static controllability in as many degrees of freedom with four, five or six coils, which corresponds to the theoretical limit, as shown in Fig. 13 for example for a magnet arrangement common to both stators with one pole pair.

[0333] The stators assumed for FIG. 13 comprise only one coil per phase. To increase the torque density of a motor, multiple coils per stator phase are often used, sometimes in conjunction with an increase in the number of rotor pole pairs. The multiple coils per phase are arranged as concentrated or distributed coils around the circumference. However, the resulting symmetry reduces the efficiency or even the controllability in Fx, Fy, Mx, and My. Often, the Mz torque density and Mz efficiency are given greater weight. For this purpose, the double-sided axial flux motor can be designed with multiple coils per phase on one side and one coil per phase on the other. This ensures good Mz efficiency with one stator and good controllability with the other.

[0334] In motors designed for lower speeds or higher drive torques, the independent coil units typically consist of several interdependent coils, connected, for example, in series or parallel. Controllability can be particularly limited if the dependent coils are evenly distributed around the circumference. Thus, most three-phase brushless DC motors, usually equipped with two independent coil units, contain several concentrated or distributed coils positioned evenly around the circumference. However, to ensure sufficient controllability in multiple degrees of freedom, the coils of an independent coil unit can be distributed over only a portion of the circumference, particularly a sector spanning an angular range of 90° or 120°. The embodiment shown in Figures 12A and 12B, in contrast to Figure 12B, is arranged differently.Figure 1A shows no coils aligned parallel to the actuated component. Instead, an alternative configuration of an axial flux motor is depicted, constructed with coils wound torodically around a magnetic return component. This is intended to illustrate that the presented methods for controllability analysis and corresponding design of specific electrical assemblies with six-dimensional controllability can be applied to all known motor configurations, especially motor configurations with an axial flux component.

[0335] The following describes various electrical systems, each comprising an electrical assembly of the type described and a control interface (not explicitly shown in the figures) by means of which the assembly, in particular the stationary component of the assembly, can be connected or is connected to a control unit, in particular the control unit mentioned above.

[0336] FIG. 14 shows a cardiac support system 500 with a blood pump 600 (rotary fluid pump) and a control unit 700 connected to the blood pump 600 by means of a driveline 710.

[0337] The blood pump 600 comprises an axial flux motor 610 with a stator 620 (stationary component) and a magnetically lubricated rotor 640 (actuatorable component) that can be rotated around a rotational axis 1400 for pumping blood. The stator 620 comprises a plurality of coil units 621.

[0338] The stator 620 is arranged on a housing 601 of the blood pump 600, which forms a cavity 604 in which the rotor 640 is received. The housing 601 includes a fluid inlet 602 and a fluid outlet 603, which can be fluidically connected to respective blood vessels and / or a heart.

[0339] The rotation axis 1400 is the main axis of a rotation of the rotor caused by the coil units 621, which drives the fluid.

[0340] The rotor 640 includes a magnet arrangement 642 which interacts with the magnetic fields generated by the coil units 621, thus forming the axial flux motor 610.

[0341] The stator 620 includes a sensor 630 (acceleration or force sensor) that can be used to minimize vibrations as described above. The blood pump 600 is an implantable blood pump (implantable VAD), although at least partially extracorporeal versions are also conceivable. The control unit 700 is an extracorporeal control unit. However, the control unit can also be fully or partially integrated into the blood pump and / or implantable with it. It is also possible for the control unit to include both components located on the blood pump and components located externally / extracorporeally.

[0342] The blood pump 600 is a centrifugal pump in the example shown, but is not limited to this. FIGS. 15A and 15B show a blood pump 800 designed as an axial pump. This comprises a rotor 820 mounted in a housing 801, with blades 822 and the magnet assembly 821. A plurality of coil units 810 are arranged on the housing 801, comprising a first group of coil units 810 and a second group of coil units 810. The coil units 810 of the first group are arranged in a first plane A-A, and the coil units 810 of the second group are arranged in a second plane CC (parallel to the first plane A-A), thus forming two parallel stators. The rotor 822 comprises two corresponding magnet assemblies configured to interact with the respective stators.In particular, six or more coil units 810 can be divided between the two levels AA and CC (for example in the combination 3+3, 3+4 or 4+4 etc.) in order to control all six degrees of freedom and thus, for example, to precisely control blade gaps and vibrations.

[0343] The coils can be arranged as in a radial flux motor, as shown in Fig. 15A. The axial offset of the group of coil units 810 relative to the magnet arrangement 821, as shown, gives the radial flux motor an additional axial flux component. The controllability of the configurations shown in Fig. 13 also applies with respect to this axial flux component. This allows, in particular by rotating the stators 810 or the magnet arrangements 821 relative to each other, controllability in six degrees of freedom with a total of six independent coil units.

[0344] The fan 100 shown in FIG. 16 can be implemented with either a first-fault-proof motor or a version with fewer phases. The fan 100 comprises two rigid components: housing 113 and rotor 114. The rotor 114 contains blades 102 for conveying fluid and a magnet assembly with motor magnets 103. These are driven by an axial flux motor consisting of a stator 104 with an optional iron core 105 and the motor magnets 103. The conveyed medium flows through openings 101 in the housing. The rotor position is measured by rotor position sensors 107, which determine either the position of the magnets 111 and 112 or the position of an optional sensor target 108. An electronic control unit 106 controls the stator 104 based on the sensor data. In principle, storage is possible without the passive magnetic bearings formed by magnets 109, 111, 112.These, however, enable more efficient operation through the use of a zero-force controller and additional emergency running properties. If the active magnetic bearing is not fully operational, the magnet pair 109 / 111 or 112 will attract the rotor, causing the rotor to make contact on one side within a catch bearing. The magnet pair on the other side then provides tilt stabilization. To improve the emergency running properties, a sliding bearing material 110 can be used as the running surface. A further advantage of a fully active magnetic bearing in fans or pumps is that the forces acting on the rotor can be measured. These forces can be determined via the actuator force or the rotor displacement, possibly corrected for gravity using an accelerometer. This allows the flow rate and, in particular, the dynamic pressure to be measured.This allows for the detection of a blockage in the intake tract or increased back pressure due to dirty cooling fins.

[0345] The fan shown in Fig. 16 can also be equipped with a motor according to Fig. 12A, 12B, 12C, or 13. The double-sided stator makes it possible to statically compensate for passive magnetic forces, such as the attractive forces between the magnet assembly and the return iron. At the same time, a separate passive magnetic bearing is unnecessary. As a result, all magnets of the rotor are also part of the magnet assembly 103. This keeps the weight of the rotor low, while good magnetic coupling to the return irons allows for efficient motor design. Both of these properties make it possible to use weaker and, in particular, more cost-effective ferrite magnets.

[0346] A reaction wheel can be designed as a modification of the fan according to FIG. 16. For this purpose, only the blades 102 are designed as a flywheel. To extend the first-fault safety to the sensors 107, 108, 111, it can be implemented with multiple or overdetermining sensors. Versions with passive magnetic bearings on both sides, single-sided magnetic bearings, or without passive magnetic bearings can also be realized. When using passive magnetic bearings, damping of the resonant frequencies is particularly important, since the damping in a vacuum is zero and can only be passively damped via lossy eddy currents or actively. The electrical assembly can be operated as a generator, while position control is simultaneously performed via the same coil units. The pump systems shown in FIGS. 14, 15A, and 16 can therefore also be used for power generation.

[0347] Using a patented electrical assembly, a mass or measuring system can be stabilized in space in up to six degrees of freedom. This is not only interesting for generating drive torque. The movement of a mass can be used, as with a reaction wheel, to compensate for vibrations of the overall system. For this purpose, an accelerometer can be attached to the stator, and its sensor data can be used to minimize stator movements via feedback. The stationary component corresponding to the rotor can be blocked or at least restricted in some degrees of freedom.

[0348] An example is shown in FIG. 17. The stabilization device 200 is designed to stabilize a reference system 201 connected to coil units of the stator (stator coils 202). For this purpose, a reaction mass 204 is suspended above the stator by springs 206. The magnets 203 in the reaction mass allow it to be moved by means of the stator coils 202 in order to exert a force on the reference system 201, which is connected to a reference sensor 205, by means of action = reaction.

[0349] A modification of the example shown in FIG. 17 is illustrated in FIG. 17. The stabilization device 200 is configured here to stabilize a sensor system, in particular an optical imaging system comprising an image sensor 210 and a lens 211. To simultaneously keep the overall mass low for a mobile system, the stator 202 does not act on a reference mass, but on a sensor carrier 212. In this case, the reference sensor 205 is rigidly connected to the motor magnets 203 via the sensor carrier 212.

[0350] As mentioned, a given coil unit 1210 can comprise one or more coils or windings. A coil unit specifically includes the coils or windings which, due to their interconnection, can only be energized together, but independently of other coil units 1210. In the case of an electric motor, such a coil unit 1210 can correspond, in particular, to a motor phase.

[0351] The stationary components 1200 described above, for example, show seven condensed coils. The corresponding actuating components 1300, which together with the respective stationary component 1200 form an efficient actuation system (such as an electric machine with stator and rotor), can, for example, have five pole pairs (see, for example, FIG. 5 / FIG. 8 and their description). However, it can also be advantageous to use significantly higher numbers of pole pairs. This can, for example, reduce iron losses in the stator if the electric machine is to be scaled up. The length of the individual magnetic circuits can thus be shortened. With higher numbers of pole pairs, the rotor and / or stator can also be made flatter.

[0352] With a higher number of pole pairs in the rotor, the number of coils per phase (i.e., coil unit) in concentrated windings or the number of segments in a distributed winding typically also increases. In conventional traction motors, the concentrated coils or segments are predominantly distributed evenly around the circumference.

[0353] Multiple concentrated coils per phase, or multiple segments of a phase in a distributed winding, can also be distributed over any angle, not just 360°. For angles smaller than 360°, controllability in six degrees of freedom is maintained. The selection property is also retained, and controllability is maintained even if only six of the phases are functional.

[0354] FIG. 19A shows a simple example of a stationary component 1200 with two lumped coils 1212a, 1212b per phase, each phase corresponding to one of the coil units 1210. The coils are arranged in the same sector 1213, corresponding to an angular range of a = 360° / Q, and wound in opposite directions (as indicated by arrows), where Q is the number of sectors, i.e., in the example shown, Q = 7. More than two coils per phase can also be implemented. The number of pole pairs of the associated rotor can be twice as large as for a stator with one lumped coil per phase.

[0355] FIG. 19B shows a possible arrangement of sectors 1213 in which the coils of the respective phases are arranged. Each arrow corresponds to one of the rotating sectors. The sectors 1213 are rotated relative to each other by α = 360° / P, where P is the number of phases. The span β of the sectors can be chosen arbitrarily, and β can be greater than α. This causes the sectors 1213 of the individual phases to overlap. The coils of the individual phases are no longer necessarily directly adjacent, but in concentrated windings can also be located between coils of other phases. In a distributed winding design, the individual segments 1213 of a phase are distributed across the sector. Here, too, as with conventional distributed windings, the segments 1213 of one phase are located between segments 1213 of other phases. The application relates, among other things, to the following aspects:

[0356] 1. Electrical assembly (1100), comprising

[0357] a stationary component (1200) comprising a plurality of coil units (1210), wherein each of the plurality of coil units (1210) is controllable to generate a variable magnetic field,

[0358] an actuable component (1300) that is movable relative to the stationary component (1200) and comprises a magnet arrangement (1310), wherein each of the plurality of coil units (1210) is configured to generate the respective variable magnetic field such that, through interaction of the respective variable magnetic field with the magnet arrangement (1310), forces and / or moments with respect to several degrees of freedom act on the actuable component (1300),

[0359] so that by controlling any combination of at most six of the plurality of coil units (1210) a position and / or movement of the actuable component (1300) relative to the stationary component (1200) can be controlled in six linearly independent degrees of freedom.

[0360] 2. Electrical assembly (1100) according to aspect 1, wherein for each given coil unit (1210) of the plurality of coil units (1210) there exists a first position of the actuable component (1300) and a second position of the actuable component (1300),

[0361] so that in the first position, a force and / or a moment can be generated by the interaction of the magnet arrangement (1310) with the variable magnetic field that can be generated by means of the specified coil unit (1210), which acts on the actuable component (1300) with respect to a first degree of freedom of the six linearly independent degrees of freedom and

[0362] so that in the second position, by interaction of the magnet arrangement (1310) with the variable magnetic field that can be generated by means of the specified coil unit (1210), a force and / or a moment can be generated which acts on the actuable component (1300) with respect to a second degree of freedom of the six linearly independent degrees of freedom.

[0363] 3. Electrical assembly (1100) according to one of the preceding aspects, wherein for each of the six specified degrees of freedom there exists at least one first coil unit (1210) of the plurality of coil units (1210) and one second coil unit (1210) of the plurality of coil units (1210), such that by interaction of the magnet arrangement (1310) with the variable magnetic field that can be generated by each of the first coil unit (1210) and the second coil unit (1210) a force and / or a moment can be generated which acts on the actuable component (1300) with respect to the specified degree of freedom.

[0364] 4. Electrical assembly (1100) according to any of the preceding aspects, wherein each of the plurality of coil units (1210) is connected such that it can be controlled independently of any other of the plurality of coil units (1210) to generate the respective variable magnetic field. 5. Electrical assembly (1100) according to any of the preceding aspects, wherein the stationary component (1200) and the actuable component (1300) form an axial flux motor or an axial flux component of an electric motor, wherein each of the plurality of coil units (1210) corresponds to a motor phase of the axial flux motor or the axial flux component.

[0365] 6. Electrical assembly (1100) according to aspect 5, wherein the plurality of coil units (1210) comprises a first group of coil units (1210) and a second group of coil units (1210), wherein the coil units (1210) of the first group are arranged in a first plane to form a first axial flux motor stator and the coil units (1210) of the second group are arranged in a second plane to form a second axial flux motor stator.

[0366] 7. Electrical assembly (1100) according to aspect 6, characterized by a design-related offset between the respective directions of action and / or coil axes of the coil units of the first axial flux motor stator and the second axial flux motor stator, in particular by a rotational offset of an arrangement of the coil units of the first axial flux motor stator relative to an arrangement of the coil units of the second axial flux motor stator and / or by a rotational offset of respective pole pairs of two magnet arrangements relative to each other and / or by an oblique magnetization of a common magnet arrangement. 8.Electrical assembly (1100) according to one of the preceding aspects, wherein the plurality of coil units (1210) comprises at most eight, in particular exactly seven or exactly six, coil units (1210), wherein preferably the plurality of coil units (1210) comprises exactly seven coil units (1210) and the magnet arrangement (1310) of the actuable component (1300) comprises exactly two, five or nine pole pairs, or wherein preferably the plurality of coil units (1210) comprises exactly eight coil units (1210) and the magnet arrangement (1310) of the actuable component (1300) comprises exactly two or six pole pairs.

[0367] 9. Electrical assembly (1100) according to aspect 6 or 7, wherein the first axial flux motor stator and the second axial flux motor stator each have exactly three coil units (1210) and the magnet arrangement has one, three or five pole pairs.

[0368] 10. Electrical assembly (1100) according to one of the preceding aspects, wherein respective magnetic components, in particular coils and / or stator teeth, of at least two coil units (1210) of the plurality of coil units (1210) are arranged asymmetrically to each other, in particular not equidistantly, and / or are dimensioned differently and / or shaped differently.

[0369] 11. Control procedure for an electrical assembly (1100) according to any of the preceding aspects, comprising:

[0370] Determining a plurality of control specifications for controlling each of the plurality of coil units (1210),

[0371] Controlling and / or regulating a position and / or movement of the actuable component (1300) relative to the stationary component (1200) by applying a signal to the plurality of coil units (1210) according to the specified plurality of control specifications.

[0372] 12. Control method according to aspect 11, wherein the majority of control specifications are determined as a control specification vector using a matrix multiplication of a commutation matrix, in particular a commutation matrix dependent on the position of the actuable component (1300), with a force-moment vector, wherein the force-moment vector is a vector with force and / or moment entries,

[0373] wherein the commutation matrix is ​​determined on the basis of a linking matrix taking into account at least one optimization criterion, for example minimizing power loss and / or maximizing controllability, wherein the linking matrix is ​​defined such that the force-torque vector is obtained by matrix multiplication from the linking matrix and the control input vector.

[0374] 13. Tax procedure according to one of aspects 11 to 12, comprising:

[0375] Determining a cross-dependency of the control and / or regulation of the position and / or movement of the actuable component (1300) relative to the stationary component (1200) with respect to a first degree of freedom from the position and / or movement of the actuable component (1300) relative to the stationary component (1200) with respect to a second degree of freedom,

[0376] Determining and / or adjusting tax requirements, taking into account specific cross-dependencies, in particular to minimize cross-dependencies.

[0377] 14. Control method according to one of aspects 11 to 13, comprising minimizing an imbalance of the actuable component (1300) with respect to a rotation relative to the stationary component (1200) about an axis of rotation by dynamically setting and / or adjusting a position and / or orientation of the axis of rotation with respect to the actuable component (1300), in particular based on sensor data acquired by means of an accelerometer and / or force sensor arranged on the actuable component (1300) and / or stationary component (1200) and / or based on minimizing at least one of the entries of the force-moment vector.

[0378] 15. Tax procedure according to one of aspects 11 to 14, comprising:

[0379] Capturing at least one measured quantity that enables the detection of a fault condition of the electrical assembly (1100), wherein the fault condition includes in particular a failure of at least one first coil unit (1210) of the plurality of coil units (1210) and / or an unintended electrical connection of at least two coil units (1210) of the plurality of coil units (1210), and

[0380] Upon detection of the fault condition: Adapting a rule for determining the plurality of control specifications such that the control and / or regulation of the position and / or movement of the actuable component (1300) relative to the stationary component (1200) is carried out using the plurality of coil units (1210) taking into account the fault condition, in particular without the first coil unit (1210).

[0381] 16. Control method according to aspect 15, wherein, after adapting the rule for determining the plurality of control specifications by controlling any combination of at most five of the plurality of coil units (1210), a position and / or movement of the actuable component (1300) relative to the stationary component (1200) is controllable by exploiting an inertia of the actuable component (1300) in a moving, in particular rotating, state of the actuable component (1300) in the six linearly independent degrees of freedom.

[0382] 17. Control unit (700), configured to control the electrical assembly (1100) according to one of aspects 1 to 10 according to the control procedure according to one of aspects 11 to 16.

[0383] 18. Electrical system, comprising

[0384] the electrical assembly (1100) according to one of aspects 1 to 10 and a control interface by means of which the stationary component (1200) can be connected or is connected to a control unit (700), in particular the control unit (700) according to aspect 17,

[0385] wherein the electrical system is designed in particular as a pumping system for a gaseous and / or liquid conveying medium.

[0386] 19. Electrical system according to aspect 18, configured as a cardiac support system and / or fontan circulation support system, wherein the stationary component (1200) is a stator of a rotary fluid pump for pumping blood, wherein the actuable component (1300) is a rotor of the rotary fluid pump,

[0387] wherein the rotor is preferably magnetically supported or supportable relative to the stator in at least one degree of freedom, in particular in all six linearly independent degrees of freedom, and in particular actively magnetically supported or supportable. List of reference numerals

[0388] 100 - fans,

[0389] 101 - Opening,

[0390] 102,822 - Shovels,

[0391] 103, 203 - Motor magnets, magnet arrangement of actuated component 104, 309, 620 - Stator,

[0392] 105 - Iron core,

[0393] 106, 700 - Control unit,

[0394] 107 - Rotor position sensor,

[0395] 108 - Sensor target,

[0396] 109, 111, 112, 203 - Magnet,

[0397] 110 - Plain bearing material,

[0398] 113, 601, 801 - Housing,

[0399] 114, 311, 640, 820 - Rotor,

[0400] 200 - Stabilization device,

[0401] 201 - Reference system,

[0402] 202 - Stator coils,

[0403] 204 - reaction mass,

[0404] 205 - Reference sensor,

[0405] 206 - Spring,

[0406] 210 - Image sensor,

[0407] 211 - Lens,

[0408] 212 - Sensor carrier,

[0409] 301 - Sensor vector,

[0410] 302 - Controller,

[0411] 303 - Target vector,

[0412] 304 - Actuator vector,

[0413] 305, 405 - Commutation,

[0414] 306 - Phase current target vector,

[0415] 307, 407 - Power amplifiers,

[0416] 308, 408 - Phase current,

[0417] 310 - Magnetic field,

[0418] 312 - Rotor position,

[0419] 313, 413 - Rotor position sensor,

[0420] 314, 414 - Commutation matrix,

[0421] 401 - Degrees of freedom target current vector,

[0422] 402 - Degrees of freedom current error vector,

[0423] 403 - Individual current controllers,

[0424] 404 - Degrees of freedom stress vector,

[0425] 406 - Phase voltage vector, 409 - Electrical stator,

[0426] 410 - magnetic stator,

[0427] 415 - Phase voltage,

[0428] 416 - Current sensors,

[0429] 417 - Linking matrix,

[0430] 418 - Transformation from fixed to rotating, 419 - Captured degree-of-freedom current vector, 420 - Magnetic flux-oriented reference frame, 421 - Stator-fixed reference frame,

[0431] 422 - Rotor dynamics,

[0432] 423 - Force & Moment (actual),

[0433] 424 - Control loop,

[0434] 425 - Regulator,

[0435] 500 - Cardiac support system,

[0436] 600, 800 blood pump,

[0437] 602 - Fluid inlet,

[0438] 603 - Fluid outlet,

[0439] 604 - Cavity,

[0440] 610 - Axial flux motor,

[0441] 621, 810, 1210 - Coil unit,

[0442] 630 - Sensor,

[0443] 642, 1310 - Magnet arrangement,

[0444] 710 - Driveline,

[0445] 1100 - electrical assembly,

[0446] 1200 - stationary component,

[0447] 1211 - Turn,

[0448] 1212a, 1212b coil

[0449] Sector 1213

[0450] 1300 - actuated component,

[0451] 1311 - Polar pair,

[0452] 1400 - Rotation axis,

[0453] 2001 - first stator,

[0454] 2002 - second stator,

[0455] 2003 - Group without stator twist, 2004 - Group with stator twist.

Claims

1. Berlin Heart GmbH 1.224 PCT / P150349PC00 Patent claims 1. Electrical assembly (1100), comprising a stationary component (1200) comprising a plurality of at least seven coil units (1210), wherein each of the plurality of coil units (1210) is controllable to generate a variable magnetic field, an actuable component (1300) that is movable relative to the stationary component (1200) and comprises a magnet arrangement (1310), wherein each of the plurality of coil units (1210) is arranged to generate the respective variable magnetic field such that, through interaction of the respective variable magnetic field with the magnet arrangement (1310), forces and / or moments with respect to several degrees of freedom act on the actuable component (1300), so that by controlling any combination of a fixed number of the plurality of coil units (1210) a position and / or movement of the actuable component (1300) relative to the stationary component (1200) can be controlled in six linearly independent degrees of freedom, wherein the fixed number is less than a total number of the plurality of coil units.

2. Electrical assembly (1100) according to claim 1, wherein the plurality of coil units (1210) comprises exactly seven or exactly eight coil units (1210).

3. Electrical assembly (1100) according to claim 2, wherein the plurality of coil units (1210) comprises exactly seven coil units (1210) and the magnet arrangement (1310) of the actuable component (1300) comprises exactly two, five or nine pole pairs, or wherein the plurality of coil units (1210) comprises exactly eight coil units (1210) and the magnet arrangement (1310) of the actuable component (1300) comprises exactly two or six pole pairs.

4. Electrical assembly (1100) according to one of the preceding claims, wherein the actuable component (1300) is arbitrarily rotatable about an axis of rotation relative to the stationary component (1200).

5. Electrical assembly (1100) according to one of the preceding claims, wherein for each predetermined coil unit (1210) of the plurality of coil units (1210) there exists a first position of the actuable component (1300) and a second position of the actuable component (1300), so that in the first position a first force is generated by the interaction of the magnet arrangement (1310) with the variable magnetic field that can be generated by means of the specified coil unit (1210). x ,y,z-and-moment, x ,y, z-vector can be generated which acts on the actuable component (1300), and so that in the second position a second force is generated by the interaction of the magnet arrangement (1310) with the variable magnetic field that can be generated by means of the specified coil unit (1210). x ,y, z -and-moment x ,y, z -vector can be generated that acts on the actuated component (1300), where the first and second force x ,y, z -and-moment x ,y, z -Vectors are linearly independent of each other.

6. Electrical assembly (1100) according to one of the preceding claims, wherein for each predetermined system of six linearly independent degrees of freedom and for each predetermined of these six degrees of freedom there exists at least one first coil unit (1210) of the plurality of coil units (1210) and a second coil unit (1210) of the plurality of coil units (1210), so that by interaction of the magnet arrangement (1310) with the variable magnetic field that can be generated by means of each of the first coil unit (1210) and the second coil unit (1210) a force and / or a moment can be generated which acts on the actuable component (1300) with respect to the given degree of freedom.

7. Electrical assembly (1100) according to one of the preceding claims, wherein the stationary component (1200) and the actuable component (1300) form an axial flux motor or an axial flux component of an electric motor, wherein each of the plurality of coil units (1210) corresponds to a motor phase of the axial flux motor or axial flux component.

8. Electrical assembly (1100) according to claim 7, wherein the plurality of coil units (1210) comprises a first group of coil units (1210) and a second group of coil units (1210), wherein the coil units (1210) of the first group are arranged in a first plane to form a first axial flux motor stator and the coil units (1210) of the second group are arranged in a second plane to form a second axial flux motor stator.

9. Electrical assembly (1100) according to claim 8, characterized by a design-related offset between the respective directions of action and / or coil axes of the coil units of the first axial flux motor stator and the second axial flux motor stator, in particular by a rotational offset of an arrangement of the coil units of the first axial flux motor stator relative to an arrangement of the coil units of the second axial flux motor stator and / or by a rotational offset of respective pole pairs of two magnet arrangements relative to each other and / or by an oblique magnetization of a common magnet arrangement. 10.Electrical assembly (1100) according to one of the preceding claims, wherein the actuable component (1300) is configured to convey a fluid, in particular by rotation relative to the stationary component (1200), and an average density of the actuable component (1300) is adapted to a density of the conveyed medium, in particular by cavities and / or mass elements arranged in and / or on the actuable component (1300).

11. Electrical assembly (1100) according to one of the preceding claims. wherein each of the plurality of coil units (1210) is assigned to a respective sector (1213) of the stationary component (1200), each of which comprises only a part, in particular an angular section, of the stationary component (1200), wherein one or more of the plurality of coil units (1210) comprise several coils which are arranged in and / or on the sector (1213) assigned to the respective coil unit (1210) and are interconnected in such a way that they can be energized only together, but independently of the other coil units of the plurality of coil units (1210).

12. Control method for an electrical assembly (1100) according to one of the preceding claims, comprising: Determining a plurality of control specifications for controlling each of the plurality of coil units (1210), Controlling and / or regulating a position and / or movement of the actuable component (1300) relative to the stationary component (1200) by applying a signal to the plurality of coil units (1210) according to the specified plurality of control specifications.

13. Control method according to claim 12, wherein the plurality of control specifications is determined as a control specification vector using a matrix multiplication of a commutation matrix, in particular a predetermined and / or position-dependent commutation matrix, with a target force-torque vector, wherein the target force-torque vector is a vector with force and / or torque entries, wherein the commutation matrix is ​​determined on the basis of a linking matrix taking into account at least one optimization criterion, for example a minimization of power loss and / or a maximization of controllability, in particular by inversion or pseudoinversion of the linking matrix, wherein the linking matrix represents a mathematical model of at least one part of the electrical assembly (1100) which determines an actual force-torque vector as a function of arbitrary control input vectors for different positions of the actuable component (1300) relative to the stationary component (1200).

14. Tax procedure according to one of claims 12 to 13, comprising: Determining a cross-dependency of the control and / or regulation of the position and / or movement of the actuable component (1300) relative to the stationary component (1200) with respect to a first degree of freedom on the position and / or movement of the actuable component (1300) relative to the stationary component (1200) with respect to a second degree of freedom, Determining and / or adjusting tax requirements, taking into account specific cross-dependencies, in particular to minimize cross-dependencies.

15. Tax procedure according to one of claims 12 to 14, comprising: Capturing at least one measured quantity that enables the detection of a fault condition of the electrical assembly (1100), wherein the fault condition includes in particular a failure of at least one first coil unit (1210) of the plurality of coil units (1210) and / or an unintended electrical connection of at least two coil units (1210) of the plurality of coil units (1210), and Upon detection of the fault condition: Adapting a rule for determining the plurality of control specifications such that the control and / or regulation of the position and / or movement of the actuable component (1300) relative to the stationary component (1200) is carried out using the plurality of coil units (1210) taking into account the fault condition, in particular without the first coil unit (1210).

16. Control method according to claim 15, wherein, after adapting the provision for determining the plurality of control parameters by controlling an arbitrary combination of at most five of the plurality of coil units (1210), a position and / or movement of the actuable component (1300) relative to the stationary component (1200) is controllable by exploiting an inertia of the actuable component (1300) in a moving, in particular rotating, state of the actuable component (1300) in the six linearly independent degrees of freedom.

17. Tax procedure according to one of claims 12 to 16, comprising: Operating the electrical assembly (1100) in a first operating state in which seven or more of the coil units (1210) are supplied with a current to generate the variable magnetic field, Capturing at least one measured quantity that enables the detection of a fault condition of the electrical assembly (1100), wherein the fault condition includes in particular a failure of at least one first coil unit (1210) of the plurality of coil units (1210) and / or an unintended electrical connection of at least two coil units (1210) of the plurality of coil units (1210), and Upon detection of the fault condition: change from the first operating state to a second operating state in which one or more of the majority of coil units (1210) are not actively supplied with current.

18. Control unit (700), configured to control the electrical assembly (1100) according to one of claims 1 to 11 according to the control method according to one of claims 12 to 17.

19. Electrical system, comprising the electrical assembly (1100) according to one of claims 1 to 11 and a control interface by means of which the stationary component (1200) can be connected or is connected to a control unit (700), in particular the control unit (700) according to claim 18, wherein the electrical system is in particular designed as a pumping system for a gaseous conveying medium and / or liquid conveying medium.

20. Electrical system according to claim 19, configured as a cardiac support system and / or fontan circulation support system, wherein the stationary component (1200) is a stator of a rotary fluid pump for pumping blood, wherein the actuable component (1300) is a rotor of the rotary fluid pump, wherein the rotor is preferably magnetically supported or mountable relative to the stator in at least one degree of freedom, in particular in all six linearly independent degrees of freedom, and in particular is actively magnetically supported or mountable.

21. Electrical assembly (1100), comprising a stationary component (1200) comprising a plurality of coil units (1210), wherein each of the plurality of coil units (1210) is controllable to generate a variable magnetic field, an actuable component (1300) that is movable relative to the stationary component (1200) and comprises a magnet arrangement (1310), wherein each of the plurality of coil units (1210) is arranged to generate the respective variable magnetic field such that, through interaction of the respective variable magnetic field with the magnet arrangement (1310), forces and / or moments with respect to several degrees of freedom act on the actuable component (1300), so that by controlling any combination of at most six of the plurality of coil units (1210) a position and / or movement of the actuable component (1300) relative to the stationary component (1200) can be controlled in six linearly independent degrees of freedom, wherein the actuable component (1300) is arbitrarily rotatable about a rotational axis relative to the stationary component (1200), wherein each of the plurality of coil units (1210) is connected in such a way that it can be controlled independently of each other of the plurality of coil units (1210) to generate the respective variable magnetic field, wherein the stationary component (1200) and the actuable component (1300) form an axial flux motor or an axial flux component of an electric motor, wherein each of the plurality of coil units (1210) corresponds to a motor phase of the axial flux motor or the axial flux component, wherein the plurality of coil units (1210) comprises a first group of coil units (1210) and a second group of coil units (1210), wherein the coil units (1210) of the first group are arranged in a first plane to form a first axial flux motor stator and the coil units (1210) of the second group are arranged in a second plane to form a second axial flux motor stator, wherein the first axial flux motor stator and the second axial flux motor stator each have exactly three coil units (1210) and the magnet arrangement has one, three or five pole pairs.

Images